Had it with Fedora!

[Ohmster]

I have been in an out of this newsgroup so many times with Fedora issues
that some of the old timers are not even taking me seriously anymore.
Well, not *all* of them. Bit Twister and even Steve are ususally willing
to roll up their sleeves and get me up and running again. Several times I
have been advised to dump Fedora because they update their entire system
every 6 months with very major changes so that even upgrades do not
really work anymore. Now xinetd is not used and something else, systemd
handles most of what xinetd used to do before. Even Bit Twister admitted
it took him several days of eyes glazed over, reading dry man pages,
trying to figure it out (sucessfully, I might add.)

I love compiz and got it to work again with unity desktop in gnome, only
to have it die again, compiz and fedora are at odds or something, they
just cannot get a stable package that works all the time. Now that comiz
is ruined again and my desktop is practically useless because of the
changes I had to make to get compiz working, Fedora 16 is out. The darned
install disc will not even boot in the Linux machine for me to upgrade,
but it will boot in a Windows machine. I tried the preinstall but it will
not work, I am stuck with an LVM drive and the preinstall setup in the
grub menu does not reflect the lvm root so the upgrade option does not
work.

For laughs, I put in the CentOS live disc and it booted up, no problems
at all. Ahhh, a stable linux distro that I am familiar with. I can even
enter the mnt directory and find all of my old stuff again! This is an
older Pentium 4 2Ghz PC with 3Gb of ram, makes a good linux machine but
does not have a SATA interface. I installed a SATA card and use a 500Gb
drive where my current system is. I have servers setup on it and they
must be setup again on the new OS. Would you recommend a new hard drive,
install Cent to the new drive, and could I then pick my stuff off the old
drive to get this box up and running perfectly, and hopefully, for a long
time without all of these stupid Fedora updates? The darned update CD
will not even boot! I give up on Fedora and finally admit I hate it.

Any tips on how to make this change to CentOS as painless as possible and
get my apache and proftpd servers up and running quickly would be
appreciated. I use the Linux as a testing server for Dreamweaver before
publishing to the web, so I also need to get my samba up and running
rather quickly as well. Is a new SATA drive the way to go and then pick
from it to make my system stable again? This time I am open to all ideas
except for flames. I have had enough of them and really, don't you
flamers have anything better to do?

Thank you all for your patience and support.

--
~Ohmster | ohmster59 /a/t/ gmail dot com

[J.O. Aho]

Ohmster wrote:

Something people always seems to forget, Fedora is RedHat's experimental
distribution where they test everything before they implement it into RedHat
Enterprise Linux (keep in ind, what happens in RHEL will happen in CentOS).

> Any tips on how to make this change to CentOS as painless as possible and
> get my apache and proftpd servers up and running quickly would be
> appreciated.

Just see to you have the same packages installed as you had in Fedora (httpd,
proftpd, php, mod_php, mysql, ...).
At least for apache, shouldn't be more than copy the config files and the web
content and start running the service, if version of proftpd isn't too much
different and you don't use really special configurations, then just copy the
config and the content should be enough.

> I use the Linux as a testing server for Dreamweaver before
> publishing to the web, so I also need to get my samba up and running
> rather quickly as well.

I recommend you remember how you got it up and running last time you had to
set it up, you can copy the files in /etc, but creating users you have to do
again (just copy stuff from /var will not work).

> Is a new SATA drive the way to go and then pick
> from it to make my system stable again?

That really depends on, if you feel confident just use the free space you may
have on the current hard drive.

--

//Aho

[Aragorn]

On Tuesday 03 January 2012 02:57, Ohmster conveyed the following to
alt.os.linux...

>
> I have been in an out of this newsgroup so many times with Fedora
> issues that some of the old timers are not even taking me seriously
> anymore. Well, not *all* of them. Bit Twister and even Steve are
> ususally willing to roll up their sleeves and get me up and running
> again. Several times I have been advised to dump Fedora because they
> update their entire system every 6 months with very major changes so
> that even upgrades do not really work anymore. Now xinetd is not used
> and something else, systemd handles most of what xinetd used to do
> before.

Actually, systemd - which is a RedHat innovation - is not a replacement
for xinetd but for the System V init system. In itself it holds
promise, but I (and several other people with me) don't like the way
things are going.

systemd and udev now both require that "/usr" is available at boot - and
thus that it is mounted before udev kicks in - because both systemd and
udev install themselves in "/usr/sbin" now, and both Kay Sievers
(developer of systemd) and Greg Kroah-Hartman (developer of udev) seem
to be pushing towards the annihilation of "/{bin,sbin,lib}" and having
everything in "/usr/{bin,sbin,lib}".

There are all sorts of problems involved with that, plus that this is
not the way UNIX was conceived. There was a reason as to why there are
"/bin", "/sbin" and "/lib" directories outside of "/usr".

> Even Bit Twister admitted it took him several days of eyes glazed
> over, reading dry man pages, trying to figure it out (sucessfully, I
> might add.)
>
> I love compiz and got it to work again with unity desktop in gnome,

> [...

With Unity? Ewww! :p

> ...] only to have it die again, compiz and fedora are at odds or

> something, they just cannot get a stable package that works all the
> time.

"Fedora" and "stable" is an oxymoron. ;-)

> Now that comiz is ruined again and my desktop is practically useless
> because of the changes I had to make to get compiz working, Fedora 16
> is out. The darned install disc will not even boot in the Linux
> machine for me to upgrade, but it will boot in a Windows machine.

I'm guessing that the CD drive in your GNU/Linux machine is still one of
the older drives that cannot handle CDs of over 640 MB in size.

> I tried the preinstall but it will not work, I am stuck with an LVM
> drive and the preinstall setup in the grub menu does not reflect the
> lvm root so the upgrade option does not work.

Hmmm... That's too vague (and ambiguous) a description for me to
comment on.

> For laughs, I put in the CentOS live disc and it booted up, no
> problems at all. Ahhh, a stable linux distro that I am familiar with.
> I can even enter the mnt directory and find all of my old stuff again!
> This is an older Pentium 4 2Ghz PC with 3Gb of ram, makes a good linux
> machine but does not have a SATA interface. I installed a SATA card
> and use a 500Gb drive where my current system is. I have servers setup
> on it and they must be setup again on the new OS.

Why are you running servers on a machine that is intended as a desktop
system? Or otherwise put, why are you using a graphical user interface
on a machine that is supposed to run servers? That's a recipe for
trouble, because most of the time, it's the proprietary video drivers
that are a stability risk, and you don't want stability issues on a
server machine.

> Would you recommend a new hard drive, install Cent to the new drive,
> and could I then pick my stuff off the old drive to get this box up
> and running perfectly, and hopefully, for a long time without all of
> these stupid Fedora updates? The darned update CD will not even boot!
> I give up on Fedora and finally admit I hate it.

The problem with this particular question is that most of the commodity
SATA drives are manufactured in South-East Asia, where there were huge
floods and where there is a lot of infrastructural damage. The
production plants making those hard disks have to be reconstructed for
most part, and given that most of their tooling and machinery was
destroyed, they have to do it all over again from scratch. By
consequence, you can expect to see a lot of faulty hard drives arriving
from South-East Asia for a while, because most of their expertise was
lost in the floods.

But either way, I would make backups of all my data - as many as
possible, but you should already have been doing that in the first place
- including the contents of "/etc", and then do a clean install of
CentOS, and set up the servers you want, and then copy back their
configuration files to the new "/etc". It might require some tweaks,
depending on the versions of the servers that you were running before
against the ones in CentOS - which is now at version 6.2, I believe.

> Any tips on how to make this change to CentOS as painless as possible
> and get my apache and proftpd servers up and running quickly would be
> appreciated.

As explained above.

> I use the Linux as a testing server for Dreamweaver before publishing
> to the web, so I also need to get my samba up and running rather
> quickly as well.

Use SWAT, the Samba administration tool. Has a web-based interface, so
you can configure it via a browser.

> Is a new SATA drive the way to go and then pick from it to make my
> system stable again?

Those are two questions, and I've given you my opinion on the first
question. The answer to the second question would be "yes". Make it a
clean install and take it from there.

> This time I am open to all ideas except for flames. I have had enough
> of them and really, don't you flamers have anything better to do?

I don't think I've ever flamed you. I was one of the people who
recommended CentOS to you. ;-)

> Thank you all for your patience and support.

You're welcome. ;-)

--
= Aragorn =
(registered GNU/Linux user #223157)

[Eef Hartman]

Aragorn <str...@telenet.be.invalid> wrote:
> There are all sorts of problems involved with that, plus that this is
> not the way UNIX was conceived. There was a reason as to why there are
> "/bin", "/sbin" and "/lib" directories outside of "/usr".

Yeah, mostly because the disks at THAT time weren't big enough to hold
both the root fs and the /usr tree (think 40 MB disks).
In the _big_ Unix world /bin and /lib have been eliminated a long time
ago, and /sbin was restricted to what was REALLY needed at boot time
(including _static built_ versions of apps out of /usr/(s)bin, as you
didn't have the shared library system up yet).
After /usr was mounted, /sbin was NOT in the PATH anymore, there was
nothing there which wasn't needed during normal operations.

PS: because of older apps, /bin and /lib often were sym-linked to
their /usr equivalents.
--
******************************************************************
** Eef Hartman, Delft University of Technology, dept. SSC/ICT **
** e-mail: E.J.M....@tudelft.nl - phone: +31-15-27 82525 **
******************************************************************

[Aragorn]

On Tuesday 03 January 2012 10:48, Eef Hartman conveyed the following to
alt.os.linux...

> Aragorn <str...@telenet.be.invalid> wrote:
>
>> There are all sorts of problems involved with that, plus that this is
>> not the way UNIX was conceived. There was a reason as to why there
>> are "/bin", "/sbin" and "/lib" directories outside of "/usr".
>
> Yeah, mostly because the disks at THAT time weren't big enough to hold
> both the root fs and the /usr tree (think 40 MB disks).
> In the _big_ Unix world /bin and /lib have been eliminated a long time
> ago, and /sbin was restricted to what was REALLY needed at boot time
> (including _static built_ versions of apps out of /usr/(s)bin, as you
> didn't have the shared library system up yet).
> After /usr was mounted, /sbin was NOT in the PATH anymore, there was
> nothing there which wasn't needed during normal operations.
>
> PS: because of older apps, /bin and /lib often were sym-linked to
> their /usr equivalents.

That is what Sievers and Kroah-Hartman are obviously aiming at - it is a
fairly hot topic on the Gentoo mailing lists, both for developers and
users. (I'm not actually subscribed to either of them, but they are
available as read-only newsgroups via gmane.org and news.eternal-
september.org carries all of them.)

But I'm not so sure on whether the big UNIX world does without those
root-level directories. I consider Solaris and AIX big enough, and to
my knowledge, neither of them does that. {Free,Net,Open}BSD doesn't do
it either.

The way I have always learned and understood it is that what's under
"/usr" pertains to the multi-user operation of the system, but the
system also still does have a single-user mode (beside the boot
sequence), and "/{bin,sbin,lib}" are there for that reason. In single-
user mode - when booted up into it, not when dropped down to it from
multi-user mode - only the root filesystem is mounted. This is useful
for correcting filesystem errors on any of the other filesystems, or
performing other types of filesystem maintenance - e.g. relocation of
certain directories/files onto other physical partitions/volumes, et al.

Furthermore, somewhat related but not unovercomeable is the fact that I
think that having "/usr" on a separate filesystem is a good idea, even
if only because I have my "/usr" filesystem mounted read-only during
normal system operation. It is of course still possible to do this with
systemd - hence "not unovercomeable" - but then "/usr" would have to be
fsck'ed and mounted from within an initramfs, or if you want to go
without an initramfs, you'd have to use an...

init=/sbin/some_script.rc

... on the kernel line in the bootloader, with the last line of that
script invoking...

exec /sbin/init

... or whatever the executable will be called if you're not using a
traditional System V init system so as to make "/sbin/init" PID 1.

In my opinion, that's all more unneeded trouble just so as to enforce
somebody's view on "how it should be", whereas the old approach "simply
worked" and made sense. In distributions like Gentoo, where you compile
your own kernel and where you may decide to go without an initramfs or
an initrd, the Sievers/Kroah-Hartman vision creates these extra
difficulties, and for no other reason than that these two people feel
that it should be that way. It's also extra problematic for
distributions like Gentoo because they are rolling release distros, so
the migration to newer versions of udev (and systemd, for those who use
that instead of Gentoo's openrc) creates even more difficulties.

[Richard Kettlewell]

Aragorn <str...@telenet.be.invalid> writes:
> In my opinion, that's all more unneeded trouble just so as to enforce
> somebody's view on "how it should be", whereas the old approach "simply
> worked" and made sense. In distributions like Gentoo, where you compile
> your own kernel and where you may decide to go without an initramfs or
> an initrd, the Sievers/Kroah-Hartman vision creates these extra
> difficulties, and for no other reason than that these two people feel
> that it should be that way. It's also extra problematic for
> distributions like Gentoo because they are rolling release distros, so
> the migration to newer versions of udev (and systemd, for those who use
> that instead of Gentoo's openrc) creates even more difficulties.

I think you have it backwards. Abandoning support for separate-/usr is
not (at heart) an aesthetic decision but (following a fixed transition
cost) a means of doing less work in the long term - it is, after all,
not something that comes for free, and its benefits are at best rather
dubious.

If you want OS integrators to continue to do the work to support it, you
will have to pay them or persuade them accordingly.

--
http://www.greenend.org.uk/rjk/

[Eef Hartman]

Aragorn <str...@telenet.be.invalid> wrote:
> But I'm not so sure on whether the big UNIX world does without those
> root-level directories. I consider Solaris and AIX big enough, and to

I was mostly a HP-UX man, and even HP-UX 10.x (1990's) already did
NOT have a /bin cq /lib anymore.
OK, in the middle 90's we mostly switched to Linux (Slackware)
and I still remember we put all of our users on tcsh as both HP-UX
as well as Slackware did have the same version available, so you
could take your profile/aliases with you from HP to PC's.

Never did all that much with Solaris (we had a few small ones, for
the Transputer cross-compiler tools didn't run on HP-UX) and never
had AIX (the only IBM we ever got was a rack-mount Opteron server
with SuSE 9.x on it).
The only other _real Unix_ we ever used was Irix (Silicon Graphics).

[Richard Kettlewell]

Eef Hartman <E.J.M....@tudelft.nl> writes:
> Aragorn <str...@telenet.be.invalid> wrote:

>> But I'm not so sure on whether the big UNIX world does without those
>> root-level directories. I consider Solaris and AIX big enough, and to
>
> I was mostly a HP-UX man, and even HP-UX 10.x (1990's) already did
> NOT have a /bin cq /lib anymore.

Indeed, but it has a statically linked /sbin/init, and it demonstrably
manages a separate /usr one way or another (at least in 11.31). Solaris
10 has a dynamically linked /sbin/init but it only uses libraries found
in /lib, so it looks like separate /usr remains in principle possible.
AIX 5.1 is the puzzler, it has /etc/init->/usr/sbin/init and a separate
/usr, at least as configured here; I suspect either overmounting or an
equivalent of early userspace mounting /usr.

--
http://www.greenend.org.uk/rjk/

[Ohmster]

"J.O. Aho" <us...@example.net> wrote in
news:9mflr3...@mid.individual.net:

> Ohmster wrote:
>
> Something people always seems to forget, Fedora is RedHat's
> experimental distribution where they test everything before they
> implement it into RedHat Enterprise Linux (keep in ind, what happens
> in RHEL will happen in CentOS).

Aho, it is so good to see you again my friend! :)

Uh, yeah, I get it. Fedora is beta. But CentOS is RHEL the free version,
isn't it? The EOL date on 6.2 is 2017. I like that a LOT better than 6
months to two years like Fedora. I am familiar with Redhat and all of
it's variants so I would like to stick with some form of distro that uses
that system. In a nutshell, what are you saying? CentOS is not good?

>> Any tips on how to make this change to CentOS as painless as possible
>> and get my apache and proftpd servers up and running quickly would be
>> appreciated.
>
> Just see to you have the same packages installed as you had in Fedora
> (httpd, proftpd, php, mod_php, mysql, ...).
> At least for apache, shouldn't be more than copy the config files and
> the web content and start running the service, if version of proftpd
> isn't too much different and you don't use really special
> configurations, then just copy the config and the content should be
> enough.

Agreed. That is why I want to go to Cent instead of slackware, debian,
ubuntu, or something else. Cent is talking about using vsftpd, I have
used it and am familiar with it so I can deal with it. ProFTPD is easier
to configure but vsftpd is not that bad.

>> I use the Linux as a testing server for Dreamweaver before
>> publishing to the web, so I also need to get my samba up and running
>> rather quickly as well.
>
> I recommend you remember how you got it up and running last time you
> had to set it up, you can copy the files in /etc, but creating users
> you have to do again (just copy stuff from /var will not work).

Agreed. I will just recreate the shares with same name, same smbusers,
etc.

>> Is a new SATA drive the way to go and then pick
>> from it to make my system stable again?
>
> That really depends on, if you feel confident just use the free space
> you may have on the current hard drive.

I do not have a lot of space on my drives now, that is why I bought the ½
gigger for and went SATA for speed, and the store shelves are getting
pretty bare in the IDE section. Plus, I am out of IDE room with two DVDRW
drives and two IDE hard disks. My 400Gb disk is full beause I used it for
backup-manager and it filled the disk. I have these backups and no clue
on how to restore them. It seems the man pages on backup-manager are
pretty sparse. I would rather wipe the 400Gb IDE (How, format?) and since
my entire 500Gb SATA drive is not full, I would like to copy everything
from it, as an image preferably, to the 400Gb drive in the event
something goes wrong. But how can I do that? cp -a for the entire disk?
What about dd? I am not really sure what dd does but so far as I know, it
is pretty powerful and can do almost anything with disk drives.

GOAL: Copy 500Gb disk to 400Gb disk as an image since part of it is ext3
or ext4 for boot, the rest is LVM for home and root. Is this a realistic
possibility?

Here is my current disk status:

[paul@paulspcworks ~]$ df -h
Filesystem Size Used Avail Use% Mounted on
rootfs 50G 16G 32G 34% /
udev 1.5G 0 1.5G 0% /dev
tmpfs 1.5G 284K 1.5G 1% /dev/shm
tmpfs 1.5G 792K 1.5G 1% /run
/dev/mapper/vg_paulspcworks-lv_root
50G 16G 32G 34% /
tmpfs 1.5G 0 1.5G 0% /sys/fs/cgroup
tmpfs 1.5G 792K 1.5G 1% /var/lock
tmpfs 1.5G 792K 1.5G 1% /var/run
tmpfs 1.5G 0 1.5G 0% /media
/dev/sdc1 184G 145G 30G 84% /mnt/200_Disk
/dev/sdb1 367G 367G 0 100% /mnt/400_Disk
/dev/sda1 485M 223M 237M 49% /boot
/dev/mapper/vg_paulspcworks-lv_home
405G 204G 181G 54% /home
[paul@paulspcworks ~]$

CentOS had no issue reading the lvm drive at all.

My Fedora 16 disc would not boot at all in this machine but booted well
in any other machine. The MD5sum was wrong, downloaded it again, it was
900Mb short. Ah hah! Made no difference, still would not boot to the
disc. I ran preinstall on the machine and now have the entire Fedora 16
setup on my machine, ready to invoke at boot from the grub menu, except
that it will not work. If I pick that menu item, I get a flashing
underscore and nothing else. The PCI SATA card changed this stock P2 2Ghz
Intel PC a lot. Now I cannot ever boot if anything is in a USB port, it
will freeze up or not boot. I tried playing with the grub menu as all the
upgrade files are there but could not get it to work. We could look at it
and fuck with it but to tell the truth, I am really, really sick of
Fedora now. Now I got F16 boot files in my boot partiton that won't work,
a full /var/preupgrade/ directory with every rpm package for the new
distro in it., so it is probably better to delete all this junk, backup
my drive otherwise, install new OS, and pick configs and files from my
old copied disk back to main root disk. Not sure how to copy a whole
disck like that though.

Just to show you how fucked up Fedora is, here are the first two lines of
my grub.conf. The default is the 2nd line to put me in F15. The 1st line
should put me right into the F16 setup but does not work. Go figure but
do not bust your brain with it, I really want to dump Fedora.

[paul@paulspcworks grub]$ sudo cat grub.conf
# grub.conf generated by anaconda
#
# Note that you do not have to rerun grub after making changes to this
file
# NOTICE: You have a /boot partition. This means that
# all kernel and initrd paths are relative to /boot/, eg.
# root (hd0,0)
# kernel /vmlinuz-version ro root=/dev/mapper/vg_paulspcworks-
lv_root
# initrd /initrd-[generic-]version.img
#boot=/dev/sda
default=1
timeout=5
splashimage=(hd0,0)/grub/splash.xpm.gz
hiddenmenu
title Upgrade to Fedora 16 (Verne)
interfaces = lo eth0 192.168.15.5/255.255.255.0 \
kernel /upgrade/vmlinuz preupgrade
repo=hd::/dev/mapper/vg_paulspcworks-lv_root/var/cache/yum/preupgrade
ks=hd:UUID=80dfb32e-640e-4bc5-83fe-7fe6164799db:/upgrade/ks.cfg
initrd /upgrade/initrd.img
title Fedora (2.6.41.4-1.fc15.i686)
root (hd0,0)
kernel /vmlinuz-2.6.41.4-1.fc15.i686 ro
root=/dev/mapper/vg_paulspcworks-lv_root
rd_LVM_LV=vg_paulspcworks/lv_root rd_LVM_LV=vg_paulspcworks/lv_swap
rd_NO_LUKS rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16
KEYTABLE=us rhgb quiet vga=791
initrd /initramfs-2.6.41.4-1.fc15.i686.img
[..]

Remember, root and var are on an lvm disk. F16 entry does not
mention /dev/mapper/ except for the repo so how will it find the setup
files in /var/cache/yum/preupgrade? All I get is a blinking cursor and it
does not even try to boot.

Thanks aho, Happy New Year.

[Ohmster]

Aragorn <str...@telenet.be.invalid> wrote in news:jduft8$374$1@dont-
email.me:

> On Tuesday 03 January 2012 02:57, Ohmster conveyed the following to
> alt.os.linux...
>

[..]

Now xinetd is not used
>> and something else, systemd handles most of what xinetd used to do
>> before.
>
> Actually, systemd - which is a RedHat innovation - is not a replacement
> for xinetd but for the System V init system. In itself it holds
> promise, but I (and several other people with me) don't like the way
> things are going.

Aragorn, another very good friend! I have to keep these replies short or
I will be at this keyboard until 3AM.

[..]

>
> There are all sorts of problems involved with that, plus that this is
> not the way UNIX was conceived. There was a reason as to why there are
> "/bin", "/sbin" and "/lib" directories outside of "/usr".

Agreed. Linux was not a Sunday project and a LOT went into getting it
right.

[..]

>>
>> I love compiz and got it to work again with unity desktop in gnome,
>> [...
>
> With Unity? Ewww! :p

No. Unity totally sucks. 3 huge icons that do very little and impossible
to find yer shit on Unity. Whoever came up with that desktop needs to be
dragged out back and taught a lesson. What a joke that stupid shit is. I
ran classic gnome with compiz. There is an option for it if you install
compiz. I had it working great and then one day, update compiz, bam! The
whole thing went down the toilet, never to return again. Now I have a
pretty much useless desktop but the servers and samba work well.

>> ...] only to have it die again, compiz and fedora are at odds or
>> something, they just cannot get a stable package that works all the
>> time.
>
> "Fedora" and "stable" is an oxymoron. ;-)

I hear ya, loud and clear. Man, I am starting to *hate* Fedora with a
passion.

>> Now that comiz is ruined again and my desktop is practically useless
>> because of the changes I had to make to get compiz working, Fedora 16
>> is out. The darned install disc will not even boot in the Linux
>> machine for me to upgrade, but it will boot in a Windows machine.
>
> I'm guessing that the CD drive in your GNU/Linux machine is still one
of
> the older drives that cannot handle CDs of over 640 MB in size.

No. Two modern DVDRW that do everything, even lightscribe. Both IDE.
Since putting in the PCI SATA controller, the machine has not behaved
"normally" since. It does boot to the SATA disk, but will not boot in any
way if anything is plugged into the one and only USB2 port on the
machine. Booting options are funky now after installing the SATA
controller. It comes up after the IDE check and quickly looks for boot
media in both CDROM drives, then moves right to Fedora on the SATA drive.
Same discs that will be skipped on this machine will work fine in any
other machine. Yet CentOS discs boot just fine. Hiren's boot DVD 14.0
also boot right up. But Fedora 16 with good MD5sum will not boot at all.

Hmmm, lot of tools in Hiren's but you get the option to work in "mini
xp" or DOS. Oh, there is Parted Magic Linux and System Rescue CD Linux on
it as well as a linux password remover for Doze machines. GParted
Partition Editor and EASEUS Partition Manager Server Edition. Plus tons
of DOS and Windows rescue and tools for each OS that you can boot to. I
always keep an up to date version of Hiren's at all times. Now that Hiren
left out the commercial stuff but left the menus in place for those that
want the real deal, you have to get one from torrents or live without the
commercial stuff.

Aho, I want to wipe my 400Gb drive, copy all the stuff from my mounted
lvm drive over to the 400 in a Linux environment and can boot to any live
disc. What is a good way to to this? I would like to get my stuff off of
the LVM to ext4 on the 400 so it is easy pickings once CentOS is
installed on the 500Gb SATA drive. How can I copy all this stuff to
another disk? cp - a on the whole disk or is there a dd option or some
other way to do it? What do you think buddy?

>> I tried the preinstall but it will not work, I am stuck with an LVM
>> drive and the preinstall setup in the grub menu does not reflect the
>> lvm root so the upgrade option does not work.
>
> Hmmm... That's too vague (and ambiguous) a description for me to
> comment on.

Not worth commenting on now, since I want Fedora gone. Preinstall is
where yum installs a preinstall program that will give you the option to
pick an upgrade version to change to, Fedora 16 Verde, It then downloads
ALL the packages you need and burries them way down in
/var/cache/yum/preinstall and puts an install Fedora 16 option on your
grub.conf. There is a vmlinuz file for it and some stage files. But it
does not work on this machine, might be because of the SATA conroller,
not sure. But now I got a couple of gigs of useless crap on my machine
that is not serving me well. Here is a section of the grub.conf file,
first entry is the install option that only gives a flashin underscore
with no booting action at all...

>> For laughs, I put in the CentOS live disc and it booted up, no
>> problems at all. Ahhh, a stable linux distro that I am familiar with.
>> I can even enter the mnt directory and find all of my old stuff again!
>> This is an older Pentium 4 2Ghz PC with 3Gb of ram, makes a good linux
>> machine but does not have a SATA interface. I installed a SATA card
>> and use a 500Gb drive where my current system is. I have servers setup
>> on it and they must be setup again on the new OS.
>
> Why are you running servers on a machine that is intended as a desktop
> system? Or otherwise put, why are you using a graphical user interface
> on a machine that is supposed to run servers? That's a recipe for
> trouble, because most of the time, it's the proprietary video drivers
> that are a stability risk, and you don't want stability issues on a
> server machine.

The servers are important but not life and death. For that stuff, I use a
commercial host. This is a learning machine where I get to play with
everything and have some fun. But the samba shares to the apache server
are good for previewing websites to customers before I upload them to
godaddy. I can use Dreamweaver, see it live on my own apache server, show
the customer, make changes, and then up it to a commercial host. Nice
setup plus FTP servers for sharing cool stuff with friends and family.

Hmmm, Hiren booted into mini xp just fine. Wow, Ghost loaded and shows
every drive in the machine for imaging. Not sure if there is a Ghost
Explorer for Linux though, although it might run on wine. DriveImage XML
works but does not show the drives. My Computer does not show the drives.
This is crazy, I need a linux solution. CentOS 6.2 Live boots very well
and shows all drives.

>> Would you recommend a new hard drive, install Cent to the new drive,
>> and could I then pick my stuff off the old drive to get this box up
>> and running perfectly, and hopefully, for a long time without all of
>> these stupid Fedora updates? The darned update CD will not even boot!
>> I give up on Fedora and finally admit I hate it.
>
> The problem with this particular question is that most of the commodity
> SATA drives are manufactured in South-East Asia, where there were huge
> floods and where there is a lot of infrastructural damage. The
> production plants making those hard disks have to be reconstructed for
> most part, and given that most of their tooling and machinery was
> destroyed, they have to do it all over again from scratch. By
> consequence, you can expect to see a lot of faulty hard drives arriving
> from South-East Asia for a while, because most of their expertise was
> lost in the floods.

Ugh!

> But either way, I would make backups of all my data - as many as
> possible, but you should already have been doing that in the first
place
> - including the contents of "/etc", and then do a clean install of
> CentOS, and set up the servers you want, and then copy back their
> configuration files to the new "/etc". It might require some tweaks,
> depending on the versions of the servers that you were running before
> against the ones in CentOS - which is now at version 6.2, I believe.
>
>> Any tips on how to make this change to CentOS as painless as possible
>> and get my apache and proftpd servers up and running quickly would be
>> appreciated.
>

[..]

>
>> This time I am open to all ideas except for flames. I have had enough
>> of them and really, don't you flamers have anything better to do?
>
> I don't think I've ever flamed you. I was one of the people who
> recommended CentOS to you. ;-)

Oh God no. You were welcomed as a dear friend and that is very sincere.
CentOS I think is really what I was looking for. I even condsidered
pirating RHEL but would really rather not going down that route. Nothing
but trouble and getting help, even in here for warez is not something
that is condoned, AT ALL. I would be exiled for doing something like
that. The Linux community is really a very decent set of dedicated and
very helpful people. And why should it be? Linux is free and if one feels
the need, they always appreciate donations and I have opened my wallet
many times for anything good that is open source. Those guys deserve and
really need the support. CentOS was a Godsend. Thank you.

>> Thank you all for your patience and support.
>
> You're welcome. ;-)
>

Happy New Year Aho.

[Bit Twister]

On Wed, 4 Jan 2012 04:14:35 +0000 (UTC), Ohmster wrote:

> on how to restore them. It seems the man pages on backup-manager are
> pretty sparse. I would rather wipe the 400Gb IDE (How, format?)

Yep, formatting will meet your requirement to clean the drive.

> and since
> my entire 500Gb SATA drive is not full, I would like to copy everything
> from it, as an image preferably, to the 400Gb drive in the event
> something goes wrong. But how can I do that? cp -a for the entire disk?

I use rsync to copy partitions.

> What about dd? I am not really sure what dd does but so far as I know, it
> is pretty powerful and can do almost anything with disk drives.

Yes, and as I misunderstand it, dd will also copy any bad blocks and
flags onto new drive. :(

> GOAL: Copy 500Gb disk to 400Gb disk as an image since part of it is ext3
> or ext4 for boot, the rest is LVM for home and root. Is this a realistic
> possibility?

My suggestion, format target drive with all needed partitions.
Install OS on new drive.
Boot a rescue cd, mount each src/dest partition and use rsync to move
everything but the old OS partition.

If trying to copy old OS. I *think* initrd.img has to be rebuilt so it
runs from new dest partition. I can not find the command to see it.
I know I use labels instead of /dev/sdxx or UUID=XXXX to help get
around that problem.

[Ohmster]

Bit Twister <BitTw...@mouse-potato.com> wrote in
news:slrnjg7p87.7...@wb.home.test:

> On Wed, 4 Jan 2012 04:14:35 +0000 (UTC), Ohmster wrote:
>
>> on how to restore them. It seems the man pages on backup-manager are
>> pretty sparse. I would rather wipe the 400Gb IDE (How, format?)
>
> Yep, formatting will meet your requirement to clean the drive.

Done. Surprised me that disk tool on CentOS Live 5.6 will show and mount
lvm drives, but CentOS Live 6.2 will not. However, the complete set of
lvm2 tools is available from a terminal so I had to use lvdisplay and
lvchange to activate the lvm drives so that the fricking disk tool would
mount them for me to have at them. I used disk tool to format the 400Gb
IDE, it was all full of backup-manager stuff that I have no clue what to
do with. Now clean and empty with ext4 fs. Nice.

>> and since
>> my entire 500Gb SATA drive is not full, I would like to copy
>> everything from it, as an image preferably, to the 400Gb drive in the
>> event something goes wrong. But how can I do that? cp -a for the
>> entire disk?
>
> I use rsync to copy partitions.

You know, I am sure that creating the partitions and using rsync would be
the way to go, but I have zero experience with rsync although this would
probably be a great time to learn it. However, I tried a different
approach. If this will not work, please advise me off of it, not to late
to change as I am now booted to CentOS Live 6.2 and have not destroyed
anything on my actual run disk so far, but to remove over 50Gb of
"preupgrade crap" that I am not going to use. Lots of empty "preupgrade"
directories in my home dir, and /var/cache/yum/ was loaded with
preinstall packages. I was careful to delete only the preinstall ones,
nothing else. These all contained rpms marked as F16. Yes, I did a no no
and used root in xwindows, got a stern warning against it but had no
choice or I would never get this done. Although I could do a lot in a
term as centos user, I could do nothing in nautilus because I did not
have an elevated nautilus window and there is no practical way of making
one in Linux that I know of. Trying to bring it up by running "nautilus"
in an su term did not work.

After activating my lvm home and filesystem partitions, I could mount
them and have at them in a root nautilus window. Nice. I opened up the
400 disk in nautilus and created a "folder" for home, boot, and
filesystem. I was tempted to show hidden files and drag all contents via
gui to the repsective 400 disk directories, I decided against it and used
a root term with cp - a to copy the entire "home" partition over to the
400 disk, home folder. Man that lvm stuff has long ass weird names when
it gets mounted. Like 24 or so characters and numbers but so far the copy
seems to be going fine. When that is finally finished tomorrow, I will
repeat for filesystem and once more for boot. This should give me a nice,
clean ext4 IDE disk from which to pluck all of my stuff and configs back
from, once the new OS is installed.

Okay, it might be down and dirty, but you guys are sysadmins, do this for
a living, have the certs and experience to get this done timely and
properly. What I am doing should work for my purposes. This is a one way
deal, I have no intentions of going back to Fedora anymore, with good
reason.

My preinstall did not work, it will not boot. My install disc, even
though it boots fine in other machines and the MD5sum is correct, will
not boot. Fedora has an entire page of bugs known in Fedora 16 and
preinstall not working, not booting with lvm and raid drives, and not
booting with other drive arrangements, is making me sick. Things that
were never even an issue before are now deal breakers in Fedora 16. This
is a joke, a Linux tinkerer's playground. Not ready for prime time, nor
will it ever be. Not with a release schedule of every 6 months. They know
that many people will have boot issues with Verne and yet, they still
release it as a production release. Are they smoking crack over at Fedora
or something? Look at this list of issues known in Fedora 16!

http://fedoraproject.org/wiki/Common_F16_bugs

No thanks. My Fedora days are gone for good. I need a stable release that
I am familiar with like CentOS so I am going for it. I am not in the
middle of any websites right now so I got enough downtime to make the
switch and then cp - all my public_html testing server stuff back, plus
my own personal stuff and still leave a backup on a different drive. I
had planned to image the drive first, then if anything went wrong, I
could "put it back the way it was" but, I do not want it back the way it
was. I want something good, clean, and stable. As long as I have my
configs and files, I am happy. Looking in the /var/cache/yum directory, I
see I still have packages from Fedora 13, 14, & 15 installed and now I am
going to 16? Whoa, time to clean house.

>> What about dd? I am not really sure what dd does but so far as I
>> know, it is pretty powerful and can do almost anything with disk
>> drives.
>
> Yes, and as I misunderstand it, dd will also copy any bad blocks and
> flags onto new drive. :(

Okay, so much for dd. Shows how much I know. Thanks Bit.

>> GOAL: Copy 500Gb disk to 400Gb disk as an image since part of it is
>> ext3 or ext4 for boot, the rest is LVM for home and root. Is this a
>> realistic possibility?
>
> My suggestion, format target drive with all needed partitions.
> Install OS on new drive.
> Boot a rescue cd, mount each src/dest partition and use rsync to move
> everything but the old OS partition.
>
> If trying to copy old OS. I *think* initrd.img has to be rebuilt so it
> runs from new dest partition. I can not find the command to see it.
> I know I use labels instead of /dev/sdxx or UUID=XXXX to help get
> around that problem.

As you can see, I took a different approach that I am pretty sure is safe
and will work. Please correct me if I am wrong as I will check back here
before wiping my working system, for what it is worth. (God, did you see
that gnome Unity desktop? Who are they kidding with that crap? OMG, so
sad. I would rather use KDE plasma instead but for some reason, kde makes
me feel like it is dated or something. I like gnome.)

Okay so far the cp -a is still cruising along. I did use this method once
to put my whole home system on a different drive and then grafted it into
the Linux filesystem. It worked like a charm and I was quite pleased with
the results. I wish that cp -a would give you some sort of progress bar
like hash marks, but it works so who is to complain. At least I did not
try to drag everything over in the gui. I just do not trust that method
at all, even with hidden files showing. All of this is being done in a
root terminal with extreme caution so it should work fine.

God it is late. I soooooooooo wanted to be in bed by now. Sonia (the
kitty) already has the bed warmed up for me, time to go.

Check back with you tomorrow or real soon. Thanks Bit, you are a real
trooper. Aragon too and all the rest of the guys that really help instead
of come here to start flame wars. You guys rock. Goodnight.

[Bit Twister]

On Wed, 4 Jan 2012 08:28:47 +0000 (UTC), Ohmster wrote:
> Bit Twister <BitTw...@mouse-potato.com> wrote in
>>

>> I use rsync to copy partitions.
>
> You know, I am sure that creating the partitions and using rsync would be
> the way to go, but I have zero experience with rsync although this would
> probably be a great time to learn it.

It is a pretty powerful tool. With --delete it will remove dest files
not in src.

An example rsync usage. I have a hot backup partition used for backup/restore
I use labeled partitions, so my fstab looks something like

LABEL=cooker / ext4 defaults 1 1
LABEL=hotbu /hotbu ext4 user,noauto,acl,relatime 1 2
LABEL=2011_0_64 /2011_0_64 ext4 user,noauto,acl,relatime 1 2

Since cooker is a running OS, I not want to backup/copy
everything so I boot another OS (LABEL=2011_0_64) to do the OS rsync.

mount /cooker
mount /hotbu
rsync -aAvx --delete /cooker/ /hotbu

If I know cooker has had a major bunch of updates since last backup,
I'll go ahead and format /hotbu so rsync does not have to check
time stamp on every dest target.

Another example, I have an account on some node (mtv) and I want to
sync a subdirectory from another node (wb) same subdirectory
and I have no desire to retype source and target directory.

cd $HOME/some_sub_dir/where/ever
rsync -aAvx --delete $US...@wb.home.test:$PWD/ $PWD

Maybe I am in a VirtualBox guest and I need a small test setup with
users from a real system. I just want to exclude a bunch of subdirs
under /accounts. Process would be:

Create new guest, attach new_CentOS.iso, boot guest, run centos installer
Reboot centos guest, log in as a user, click up a terminal

$ sudo -i
# echo 192.168.1.132 wb.home.test wb >> /etc/fstab

# rsync -aAvx ro...@wb.home.test:/local/bin/ /local/bin

# rsync -aAvx --exclude-from=/local/bin/rsync.excludes ro...@wb.home.test:/accounts/ /accounts

# exit
$

[J G Miller]

On Wednesday, January 4th, 2012 at 08:28:47h +0000, Ohmster craved:

> I wish that cp -a would give you some sort of progress bar
> like hash marks

Which is why you should use rsync with the --progress parameter.

Also, if for some reason cp -a fails, then you have to start
copying from the beginning all over again, whereas rsync will
only start copying the last file which failed to be copied over.

rsync -avvz --progress "dir_with_subdirs_to_be_copied" "destination"

It is important however not to include the trailing / on the directory names
because then the behavior is slightly different to that which you probably
want.

[J G Miller]

On Wednesday, January 4th, 2012, at 05:00:55h +0000, Ohmster wrote:

> Two modern DVDRW that do everything, even lightscribe.

DVD-RAM as well?

[Aragorn]

On Tuesday 03 January 2012 14:55, Richard Kettlewell conveyed the
following to alt.os.linux...

> [...]

> Solaris 10 has a dynamically linked /sbin/init but it only uses
> libraries found in /lib, so it looks like separate /usr remains in

> principle possible. [...]

If I recall correctly, then Solaris has its "/usr", "/var" and "/home"
all under "/export", with symlinks to the root level directories or vice
versa. So depending on your preferences, you could have the
aforementioned three directories as root level directories/mountpoints
and place symlinks to them under "/export", or do it the other way
around, i.e. have the mountpoints themselves under "/export" and let the
root-mevel directories be symbolic links to them.

Either way, having a separate "/usr" should be kept possible for the
sake of mounting it read-only and/or exporting it to other machines via
NFS. Not supporting this any longer would be The Microsoft Way (tm),
which is by definition always synonymous to The Wrong Way (tm).

In a way, this whole udev/systemd "innovation" reminds me of GRUB v2:
i.e. it adds needless complexity for no other gain than the comfort of
its developers.

Your mileage may vary. ;-)

[Aragorn]

On Wednesday 04 January 2012 06:00, Ohmster conveyed the following to

alt.os.linux...

> Aragorn <str...@telenet.be.invalid> wrote in news:jduft8$374$1@dont-
> email.me:
>
>> On Tuesday 03 January 2012 02:57, Ohmster conveyed the following to
>> alt.os.linux...
>>

>>> Now xinetd is not used and something else, systemd handles most
>>> of what xinetd used to do before.
>>
>> Actually, systemd - which is a RedHat innovation - is not a
>> replacement for xinetd but for the System V init system. In itself
>> it holds promise, but I (and several other people with me) don't like
>> the way things are going.
>
> Aragorn, another very good friend! I have to keep these replies short
> or I will be at this keyboard until 3AM.

Yeah, I know you like talking. :p No problem, buddy. ;-)

> [..]
>>
>> There are all sorts of problems involved with that, plus that this is
>> not the way UNIX was conceived. There was a reason as to why there
>> are "/bin", "/sbin" and "/lib" directories outside of "/usr".
>
> Agreed. Linux was not a Sunday project and a LOT went into getting it
> right.

Well, I was talking of the original AT&T and BSD Unix versions and of
the FHS 2.3, but GNU/Linux sits in the same ballpark. ;-)

> [..]
>>>
>>> I love compiz and got it to work again with unity desktop in gnome,
>>> [...
>>
>> With Unity? Ewww! :p
>
> No. Unity totally sucks. 3 huge icons that do very little and
> impossible to find yer shit on Unity. Whoever came up with that
> desktop needs to be dragged out back and taught a lesson.

That would be Canonical, or perhaps even Shuttleworth himself. ;-)

> What a joke that stupid shit is.

It was initially designed as a GUI for netbooks and tablets, but then
Shuttleworth decided to make it the default GUI for desktops as well in
Ubuntu proper. (Other desktop variants do of course still exist, i.e.
Kubuntu with KDE, Xubuntu with XFCE and Lubuntu with LXDE.)

> I ran classic gnome with compiz. There is an option for it if you
> install compiz. I had it working great and then one day, update
> compiz, bam! The whole thing went down the toilet, never to return
> again. Now I have a pretty much useless desktop but the servers and
> samba work well.

Well, I've only briefly played around with Compiz Fusion. I use KDE 4
here, which has its own compositing window manager, albeit that in this
installation - Mageia 1 - I lost all 3D eyecandy after an update of the
proprietary nVidia driver, and nouveau doesn't do anything fancy here
either. So at the moment I'm stuck with 2D only.

>>> ...] only to have it die again, compiz and fedora are at odds or
>>> something, they just cannot get a stable package that works all the
>>> time.
>>
>> "Fedora" and "stable" is an oxymoron. ;-)
>
> I hear ya, loud and clear. Man, I am starting to *hate* Fedora with a
> passion.

Well, somebody's got to test the software, and that's what Fedora is
for: for testing. Once it hits stable status it will be incorporated in
RedHat/CentOS.

> Aho, [...

No, I am Aragorn, not J.O. Aho. ;-)

> ...] I want to wipe my 400Gb drive, copy all the stuff from my mounted

> lvm drive over to the 400 in a Linux environment and can boot to any
> live disc. What is a good way to to this? I would like to get my stuff
> off of the LVM to ext4 on the 400 so it is easy pickings once CentOS
> is installed on the 500Gb SATA drive. How can I copy all this stuff to
> another disk? cp - a on the whole disk or is there a dd option or some
> other way to do it? What do you think buddy?

Well, Bit Twister gave you some good advice on that. rsync will cover
your needs. And dd does indeed copy over the bad blocks too, because it
does a block by block copy. Wouldn't be wise either, because of
potential issues with UUIDs and LABELs.

> Here is a section of the grub.conf file, first entry is the install
> option that only gives a flashin underscore with no booting action at
> all...

That means that it's hanging somewhere, but the CPU is not halted - the
cursor would not be blinking if it were.

>>> For laughs, I put in the CentOS live disc and it booted up, no
>>> problems at all. Ahhh, a stable linux distro that I am familiar
>>> with. I can even enter the mnt directory and find all of my old
>>> stuff again! This is an older Pentium 4 2Ghz PC with 3Gb of ram,
>>> makes a good linux machine but does not have a SATA interface. I
>>> installed a SATA card and use a 500Gb drive where my current system
>>> is. I have servers setup on it and they must be setup again on the
>>> new OS.
>>
>> Why are you running servers on a machine that is intended as a
>> desktop system? Or otherwise put, why are you using a graphical user
>> interface on a machine that is supposed to run servers? That's a
>> recipe for trouble, because most of the time, it's the proprietary
>> video drivers that are a stability risk, and you don't want stability
>> issues on a server machine.
>
> The servers are important but not life and death. For that stuff, I
> use a commercial host. This is a learning machine where I get to play
> with everything and have some fun. But the samba shares to the apache
> server are good for previewing websites to customers before I upload
> them to godaddy. I can use Dreamweaver, see it live on my own apache
> server, show the customer, make changes, and then up it to a
> commercial host. Nice setup plus FTP servers for sharing cool stuff
> with friends and family.

Oh okay, I understand now.

>>> This time I am open to all ideas except for flames. I have had
>>> enough of them and really, don't you flamers have anything better to
>>> do?
>>
>> I don't think I've ever flamed you. I was one of the people who
>> recommended CentOS to you. ;-)
>
> Oh God no. You were welcomed as a dear friend and that is very
> sincere. CentOS I think is really what I was looking for. I even
> condsidered pirating RHEL but would really rather not going down that
> route.

There's absolutely no need for that. That's why CentOS exists.
Whatever you can do with RedHat proper, you can do with CentOS.

>>> Thank you all for your patience and support.
>>
>> You're welcome. ;-)
>
> Happy New Year Aho.

Thank you, and I wish you the same, but I still am Aragorn, not J.O.
Aho. :p

[Richard Kettlewell]

Aragorn <str...@telenet.be.invalid> writes:
> Richard Kettlewell conveyed the following to alt.os.linux...

>> Solaris 10 has a dynamically linked /sbin/init but it only uses
>> libraries found in /lib, so it looks like separate /usr remains in
>> principle possible. [...]
>
> If I recall correctly, then Solaris has its "/usr", "/var" and "/home"
> all under "/export", with symlinks to the root level directories or
> vice versa.

You don't recall correctly.

> Either way, having a separate "/usr" should be kept possible for the
> sake of mounting it read-only and/or exporting it to other machines
> via NFS. Not supporting this any longer would be The Microsoft Way
> (tm), which is by definition always synonymous to The Wrong Way (tm).

Remote /usr doesn't make any sense on a packagized operating system.
What is read-only /usr supposed to be for? (Don't say "security", as
that's obviously nonsense.)

> In a way, this whole udev/systemd "innovation" reminds me of GRUB v2:
> i.e. it adds needless complexity for no other gain than the comfort of
> its developers.

Don't use it, then.

--
http://www.greenend.org.uk/rjk/

[Aragorn]

On Wednesday 04 January 2012 17:21, Richard Kettlewell conveyed the

following to alt.os.linux...

> Aragorn <str...@telenet.be.invalid> writes:
>> Richard Kettlewell conveyed the following to alt.os.linux...
>
>>> Solaris 10 has a dynamically linked /sbin/init but it only uses
>>> libraries found in /lib, so it looks like separate /usr remains in
>>> principle possible. [...]
>>
>> If I recall correctly, then Solaris has its "/usr", "/var" and
>> "/home" all under "/export", with symlinks to the root level
>> directories or vice versa.
>
> You don't recall correctly.

Well, the OpenSolaris docs I have on my system say otherwise, so...

>> Either way, having a separate "/usr" should be kept possible for the
>> sake of mounting it read-only and/or exporting it to other machines
>> via NFS. Not supporting this any longer would be The Microsoft Way
>> (tm), which is by definition always synonymous to The Wrong Way (tm).
>
> Remote /usr doesn't make any sense on a packagized operating system.

Do you consider Gentoo to be "packaged"? It's installed from sources in
a chroot environment, but with an automated compilation, build and
installation of the sources.

> What is read-only /usr supposed to be for? (Don't say "security", as
> that's obviously nonsense.)

It /is/ usuable as an extra security measure, forming an extra hurdle
for anyone trying to install a rootkit, and one that they're usually not
even contemplating, so it keeps them busy for a while or might even
discourage them. Security by obscurity, I know.

It is however recommended to keep "/usr" separate because of filesystem
integrity protection - "/usr" contains nothing but static files anyway,
while the root filesystem is mostly static, but not entirely - and if
the filesystem is mounted read-only then it is even better protected
against filesystem corruption due to system stability issues.

>> In a way, this whole udev/systemd "innovation" reminds me of GRUB v2:
>> i.e. it adds needless complexity for no other gain than the comfort
>> of its developers.
>
> Don't use it, then.

If there were an easy way of getting out from underneath it, I would.
So far only openSUSE seems to offer an easy way to replace systemd by a
traditional SysV init. Getting rid of udev is a lot harder, because
lots of stuff depends on it nowadays.

[Richard Kettlewell]

Aragorn <str...@telenet.be.invalid> writes:
> Richard Kettlewell conveyed the

>> What is read-only /usr supposed to be for? (Don't say "security", as
>> that's obviously nonsense.)
>
> It /is/ usuable as an extra security measure, forming an extra hurdle
> for anyone trying to install a rootkit, and one that they're usually not
> even contemplating, so it keeps them busy for a while or might even
> discourage them. Security by obscurity, I know.

i.e. obviously nonsense. (i) if / is writable then the rootkit can go
there. (ii) if someone is in a position to install a rootkit then
they're generally also in a position to remount /usr rw.

--
http://www.greenend.org.uk/rjk/

[John Hasler]

Richard Kettlewell writes:
> (i) if / is writable then the rootkit can go there. (ii) if someone
> is in a position to install a rootkit then they're generally also in a
> position to remount /usr rw.

If the rootkit is being installed by some sort of malware rather than by
a person it is unlikely to be prepared for read-only partitions. Same
goes for noexec /home.

--
John Hasler
jha...@newsguy.com
Dancing Horse Hill
Elmwood, WI USA

[J.O. Aho]

(i) only if the rootkit implementer is a person, most scripts don't have a
check for what is writeable, they either get successes or failure.
A human user could of course mount a ramdisk and have the rootkit work from there.

(ii) That is a bit more difficult to do if the /usr is a NFS share as Aragorn
mentioned, as it's more likely that the exporting machine has the read only
for the client in question.


--

//Aho

[Ohmster]

Bit Twister <BitTw...@mouse-potato.com> wrote in

news:slrnjg8nf3.f...@wb.home.test:

> On Wed, 4 Jan 2012 08:28:47 +0000 (UTC), Ohmster wrote:
>> Bit Twister <BitTw...@mouse-potato.com> wrote in
>>>
>>> I use rsync to copy partitions.
>>
>> You know, I am sure that creating the partitions and using rsync
>> would be the way to go, but I have zero experience with rsync
>> although this would probably be a great time to learn it.
>
> It is a pretty powerful tool. With --delete it will remove dest files
> not in src.
>
> An example rsync usage. I have a hot backup partition used for
> backup/restore I use labeled partitions, so my fstab looks something
> like

[..]

Wow, rocket science! Dude, let me grab my lab coat, pocket protector, and
sliderule and I will be right with you. ;>)

BT, as ususal, you've outdone yourselve. This is a magnificent tutorial
on how to use rsync to backup and restore file systems. At this point, cp
-a has pretty much given me what I want and I am satisifed with it.
Aragorn or aho wanted to know what the hell I am doing running servers on
a box with Xwindows and all the eye candy. My answer was simple, I use
this as a testbed and learning toy, yet the servers are not mission
critical but quite useful and necessary. If this were a serious website
or mission critical stuff, I would get some hosting space at godaddy or
somewhere to keep it secure and available, 24/7. I use Dreamweaver to
make sites, samba to access them on the testbed server, check them live
and online, give the customer previews that way, and when finalized,
upload the site to a real hosting site with domain and DNS. The ftp
servers are for me to have remote access and to share cool stuff with
friends and family. With a comcast cable account, even 2nd from the
bottoms service, rocks like hell. I get downstream up to 99Mbps (On a
good torrent with mulitple feeds) and upload speed of several mpps. For
me, that is a super cool Linux playground, and a useful and functional
one.

I stated my intentions to drop fedora like a hot rock and my plan to
backup what I have and put it back to use in a similar, but stable
operating system. CentOS looks like the way to go, RHEL was one thought
but it is expensive. The idea to pirate RHEL makes me feel like crap, I
just cannot do that to the Linux community and then come here and ask for
help. Then I heard about CentOS and my prayers were answered. I looked at
the EOL date on Cent 6.2 and see it is 2017! Wow, after getting burned on
redhat 9 and letting it go way past EOL and having the script kiddies get
in via apache, I desperately want to get out of this update every 6
months stuff. Security updates, fine. Some new features, great, but the
entire OS from scratch, every 6 months? Uh no. Fedora is for
experimenters and that is pretty cool but I rely on this machine too much
for that.

Since no one strongly objected to my cp-a, I am almost done with it and
am satisfied that I have not lost any of my important stuff. Oh, look at
that, it is done! So that is it then. I am going to put the CentOS 6.2
install set in and let 'er rip. The configs should be very close to the
same so that should not be a major issue. I know how to work the os and
the package system. I am good to go.

However, I do so want to get a much better backup scheme than backup-
manager. I used it and ended up with a full 400Gb disk of stuff I have no
clue what to do with. So, once I am up and running, I am going to learn
and implement rsync and your tutorial will be in high demand. I am going
to be truthful, bit. I read your page and have not a genuine clue on how
the hell you did all that stuff and how it works. It is well done and a
good project to study and implement. But I will have to consider it a
lesson and learn it over time. Thank you.

Now JG Miller did a real nice thing by giving me a simple command line
that I can work with that will show progress and that would have been
most useful had I not already backed up what I need. You guys gave me
great advice, never condescending, always helpful and upbeat. That is why
I come here and I pay it back when I can answer something to free you
guys up for the big jobs. I met so many terrific people here that have
helped so much. Cousin Stanely with his python scripts, bit twister with
everything imaginable, aho who always chimes in with something useful,
aragon who does the same, sinner who helps with usenet stuff and shark
(RIP) who was such an advocate for genuine usenet and no google groups.
Steve Ackerman who actually rolled up his sleeves to pull out a redhat
box and solve the issue of nvidia drivers not working anymore because
"kernel headers" were never found anymore. That dude put a LOT of work
into figuring this thing out and giving me what I want. Now Steve, as
brilliant as he is, has a short fuse for people who make the same mistake
over again and I am guilty of that. I get help and come back two years
later with the same issue. Steve has never been mean but he does not have
a lot of patience for stuff like that and who can blame him? He has my
respect and is not likely to lose it anytime soon, just like all the rest
of you.

I do not know if I will ever get my compiz back the way I want it. I
might achieve it but one fellow in this thread did it and after one
compiz update, lost it all and is now stuck with 2D again. I don't get
it. As cool as compiz is, it is like Fedora, here today, gone tomorrow,
may show up again one day, who knows. Compiz is the best and coolest 3D
desktop that I have ever seen but they just cannot seem to make it stick,
one team works on one thing, the other takes a different route, and
often, the twain shall never meet. I am afriad to work on compiz again
because it will take a lot of work to make it go, and be nice, and one
day, up2date will wipe it out and there will be no way to fix it, and now
my system is poluted with stuff that no longer works and probably won't
work again. Once that happens, my desktop is ruined and never works quite
right again. KDE plasma had a pretty good 3D element but they did not go
far enough with it, why not just incorporate compiz with all the bells
and whistles? kde is probably really good, dependable, and stable, but I
just do not like it, makes me feel like I am trapped into a dated system
or something. Gnome is always cool and has room for new things.

Oh dear, I am going on way to long here. Time to go. Thanks for the rsync
tutorial bit. I will need time to study it and probably test it in short
chunks, proving it works, bit by bit, and finally pulling it all together
like you have done with your cooker demonstration. Mom is here, taking me
out for a birthday rib dinner. Who can say no to that? See ya and thanks.
TTYL.

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je1rnt$d0$1...@dont-email.me:

JG, this is perfect. Exactly the kind of stuff I need. Thank you and Happy
New Year.

[J G Miller]

On Wednesday, January 4th, 2012 at 22:17:07h +0000, Ohmster wrote:

> JG, this is perfect. Exactly the kind of stuff I need. Thank you and
> Happy New Year.

Thank you for the good wishes.

May you have a Happy "Fedoraless" 2012 ;)

[Bit Twister]

On Wed, 4 Jan 2012 22:16:19 +0000 (UTC), Ohmster wrote:


> Since no one strongly objected to my cp -a, I am almost done with it and

If you have a system where acl's are used, you will be in for a bit of
a surprise. :( They will be lost, and probably will not run on
destination partition. :-(

Of course you do not get the feature of removing files on destination which
are no longer on source partition that the --delete would provide.

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je1rsu$d0$2...@dont-email.me:

> On Wednesday, January 4th, 2012, at 05:00:55h +0000, Ohmster wrote:
>
>> Two modern DVDRW that do everything, even lightscribe.
>
> DVD-RAM as well?
>

I am not sure. I have never used or seen a need for DVD-RAM. How is it
used? These burners are within the 2 year mark, I think. When I get a
stable distro installed, I might find a tool to test for it. I know that
Nero has an excellent drive info tool for doze, but for some reason, the
only Nero Linux I can find is version 4 and I don't believe it has the
full suite of tools with it. At least version 4 will burn DVD discs.

Right now I am running CentOS 6.2 live but there are surprisingly few
tools for a DVD live release. Disk Utility does ID the drives:

Asus DRW-1814BL (IDE)
Lite-On DVDRW SOHW-8125 (IDE)

You know, Disk Utility is a surprisingly useful GUI tool for drive
managment, lot of information from it plus some SMART tests and tools.
Cent 5.6 will recognize and mount an LVM drive, Cent 6.2 won't. (Live
Discs) To it's credit, 6.2 does have the full set of LVM2 CLI tools
available so I can use pvdisplay and pvchange to activate the partitions,
at which point Disk Utility will mount them. Strange, they seem to be
going backwards with such a fine GIU utility. Ah well, CLI managment is
really where it is at and I do need to get better at it. I like to use
PuTTY to ssh into the machine and do CLI work on it. That plus Webmin
makes a surprisingly useful set of tools to manage even a headless
machine, although with the machine in the same room as I am in, I can use
X with all desktop utilities on it if I want.

I have to force myself to dismount these lvm drives prior to rebooting.
Who knows what damage could occur if I shut down with mounted lvm drives.
I did screw up my lvm system once and it took 10 days for me to figure
out how to "bring it back". I used Webmin to "create a new partition" on
my system drive, for what reason, I cannot remember, but that totally
hosed the system. I did read a dd command somewhere that would capture
the first few lines on the disk to a file but was warned that every time
I tried to access the drive, this important structure information would
be pushed out of the buffer and be gone. I got enough of it saved to file
that I could create a skeleton rescue version of the drive and use the
restore option to repair it. It worked! Since then, I am really scared of
lvm drives and do not see the purpose of creating lvm system drives for
every install on a redhat based os. Why not use that only for multiple
drives to make one big one, what is the point of not using ext3 or 4 and
leaving it readable as it is? Who knows why they do this, but that is the
default redhat, Fedora, and maybe RHEL and CentOS too. I will find out
soon enough. Cheers and thank you.

[Ohmster]

Aragorn <str...@telenet.be.invalid> wrote in news:je1tu6$h94$1@dont-
email.me:

> On Wednesday 04 January 2012 06:00, Ohmster conveyed the following to
> alt.os.linux...
>

[..]

>> Aho, [...
>
> No, I am Aragorn, not J.O. Aho. ;-)

It was past 3:00 in the morning, I was running cp -a and being critical of
ANY syntax error that might screw it up, and must beg your pardon. You are
a good dude.

[..]

>>>> Thank you all for your patience and support.
>>>
>>> You're welcome. ;-)
>>
>> Happy New Year Aho.
>
> Thank you, and I wish you the same, but I still am Aragorn, not J.O.
> Aho. :p
>

You deserve and will get a better answer than this, but I got invited to a
birthday rib dinner and do not want to miss it. I will come back to chat
with you later on. You make some darned good sense and I am glad to see
that I am not the only one with an interest in a 3D desktop.

Cheers and TTYS, Aragorn.

[unruh]

On 2012-01-04, Ohmster <ro...@dev.nul.invalid> wrote:
> Bit Twister <BitTw...@mouse-potato.com> wrote in
> news:slrnjg8nf3.f...@wb.home.test:
>
>> On Wed, 4 Jan 2012 08:28:47 +0000 (UTC), Ohmster wrote:
>>> Bit Twister <BitTw...@mouse-potato.com> wrote in
>>>>

> one.
>
> I stated my intentions to drop fedora like a hot rock and my plan to
> backup what I have and put it back to use in a similar, but stable
> operating system. CentOS looks like the way to go, RHEL was one thought
> but it is expensive. The idea to pirate RHEL makes me feel like crap, I

Uh, wht is expensive under Redhat is not the release, but the support
contract. RHEL is a GPL distribution. Ie, it is completely legal for
anyone to copy it and use it. That is what Linux is all about. It is NOT
pirating. This is exactly what Centos does. It takes Redhat and copies
it.

> just cannot do that to the Linux community and then come here and ask for

Do what? Use linux exactly as it is supposed to be used?

> help. Then I heard about CentOS and my prayers were answered. I looked at
> the EOL date on Cent 6.2 and see it is 2017! Wow, after getting burned on
> redhat 9 and letting it go way past EOL and having the script kiddies get
> in via apache, I desperately want to get out of this update every 6
> months stuff. Security updates, fine. Some new features, great, but the
> entire OS from scratch, every 6 months? Uh no. Fedora is for
> experimenters and that is pretty cool but I rely on this machine too much
> for that.
>
> Since no one strongly objected to my cp-a, I am almost done with it and
> am satisfied that I have not lost any of my important stuff. Oh, look at

Well, I would not bet on it. One of the features of rsync is that it
does a hash of the two files to ensure that they really really are the
same. It is also much faster and does not copy stuff that has already
been copied. Start using rsync.

[J G Miller]

On Thursday, January 5th, 2012, at 00:53:39h +0000, Unruh wrote:

> RHEL is a GPL distribution. Ie, it is completely legal for
> anyone to copy it and use it.

Except for the Red Hat artwork and branding, and any restrictively
copyrighted software it contains.

> That is what Linux is all about.

Strictly speaking this is what the GPL is all about.

> This is exactly what Centos does. It takes Redhat and copies it.

And removes all the Red Hat artwork and branding and any restrictively
copyrighted software it contains.

In fact, CentOS removes some Red Hat packages, modifies a number of
packages, and adds three Centos identity/indexing packages of its
own.

<http://wiki.centos.ORG/Manuals/ReleaseNotes/CentOS6.2#head-3e5a9388fbf8c845786dc6c456dce210f9e41a96>

[J G Miller]

On Wednesday, January 4th, 2012, at 22:36:09h +0000, Ohmster wrote:

> J G Miller <mil...@yoyo.ORG> wrote in news:je1rsu$d0$2...@dont-email.me:
>
>> On Wednesday, January 4th, 2012, at 05:00:55h +0000, Ohmster wrote:
>>
>>> Two modern DVDRW that do everything, even lightscribe.
>>
>> DVD-RAM as well?
>>
>>
> I am not sure.

That would tend to make your statement irresponsible then.

> I have never used or seen a need for DVD-RAM. How is it used?

Well, you remember the horrors of packet writing to CD or DVD?

DVD-RAM is using a DVD just like a floppy diskette or hard disk or USB memory
stick without any of the frustrations of the inadequacies of packet burning.

> but for some reason, the only Nero Linux I can find is version 4 and
> I don't believe it has the full suite of tools with it.

Why would you not use brasero or k3b for working with CDs and DVDs?

With DVD-RAM, you just create a UDF file system on the DVD-RAM disc
and then mount it and use as you would with a file system on a floppy
diskette, hard disk, or USB memory stick.

<http://ubuntuforums.org/showthread.php?t=1028192>

After mounting the file system, you can, if you are GUI inclined,
use drag and drop from your favorite file manager.

> Asus DRW-1814BL (IDE)

So you bought something and do not know what it does?

QUOTE

The ASUS DVD Burner enables users to burn any
DVD+R/-R, DVD+RW/-RW or *DVD-RAM* disc


UNQUOTE

> Lite-On DVDRW SOHW-8125 (IDE)

Are you absolutely certain that is 8125 and not 812S?

You stated "These burners are within the 2 year mark, I think."

I cannot find any reference to an 8125 and the 812S model
is from 2004, which is over 7 years old.

<http://www.hardwarereview.net/Reviews/Dual%20Format%20DVD%20Burners/DualFormatBurnerRoundup.htm>

And the ASUS DRW-1814BL is a 2007 model so that makes it 4 years old at least,
twice the age that you were under the impression it was.

[Gordon]

On 2012-01-03, Ohmster <ro...@dev.nul.invalid> wrote:
>
> I have been in an out of this newsgroup so many times with Fedora issues
> that some of the old timers are not even taking me seriously anymore.
> Well, not *all* of them. Bit Twister and even Steve are ususally willing
> to roll up their sleeves and get me up and running again. Several times I
> have been advised to dump Fedora because they update their entire system
> every 6 months with very major changes so that even upgrades do not
> really work anymore. Now xinetd is not used and something else, systemd
> handles most of what xinetd used to do before. Even Bit Twister admitted
> it took him several days of eyes glazed over, reading dry man pages,
> trying to figure it out (sucessfully, I might add.)

Fedora is the cutting edge test bed for Red Hat. It is not a "stable"
system, and the developers do stuff if they can do it. Fair enough test it
out and if it crashes and burns then they try something else.

Fedora should not be used in a production/serious environment

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je32f5$7fu$2...@dont-email.me:

*ahem*

Mr. Miller, you really did catch me out there. Were you ever a
schoolteacher, perhaps?

> On Wednesday, January 4th, 2012, at 22:36:09h +0000, Ohmster wrote:
>
>> J G Miller <mil...@yoyo.ORG> wrote in news:je1rsu$d0$2...@dont-email.me:
>>
>>> On Wednesday, January 4th, 2012, at 05:00:55h +0000, Ohmster wrote:
>>>
>>>> Two modern DVDRW that do everything, even lightscribe.
>>>
>>> DVD-RAM as well?
>>>
>>>
>> I am not sure.
>
> That would tend to make your statement irresponsible then.

Okay, I will wear it, if you insist.

>
>> I have never used or seen a need for DVD-RAM. How is it used?
>
> Well, you remember the horrors of packet writing to CD or DVD?

Ah, I am trying to think. Not really, but I do remember the very early
days when coasters were pretty much expected and sometimes half of your
blank media ended up in the trash for one reason or the other. Me and a
total computer geek friend that I met on the Internet when it first
became available, (He would play Doom with me if I would play Indycar
with him at our own LAN parties.) go back to the days when a gigabyte
hard drive was a dream that nobody ever got to realize, until one day, he
came over with a sealed silver bag. He did it, he actually put his
fingers on a real one gigabyte hard drive. We were amazed. RAM cost close
to $300 for a megabyte and I actually ponied up for it. I wanted it that
badly. We scoured the net for free software and games and amassed quite a
collection. So much so that storing it became a real issue. Gary could
use a hex editor to "hack" shareware by eliminating the help file, which
at the time, was the nag screen prompting you to purchase it. Not all of
this software was worth the few dollars that the developer asked for so
this hacking was not thought of as a terrible thing to do at the time.
Yes, we did pay for worthy software that we used and even donated up to
five dollars for the really good open source stuff. Gary had a friend who
actually had a real CD burner on it and not only that, he was willing to
come to my house, bring his rig, go on the LAN, and burn some of our
stuff onto CDs to immortalize forever. Wow. He did and he warned us that
burning CDs was dicey and coasters were a frequent consequense of the
ordeal. We forged ahead and burned our own CDs and yes, we made a few
coasters in the process. But we did it. We went from piles of floppy
disks to a few CDs and were astounded by the entire affair. We were Gods.

I never used a DVD-RAM before, although many of the drives I have
purchased could create them. I never saw the need for one. Just burn a
standard, one shot CD and save your stuff forever. At that time, Roxio
had this very annoying "feature" to burn to CD at any time by dragging
files to a desktop icon at any time. I never used this feature but that
driver caused a LOT of headaches with many Windows sytems. So much so
that Roxio finally took that dammed thing out and it bothered no one
anymore.

Windows 7 Explorer had a nice feature to burn a "Live CD" that would
create a file system on a CD and you could drag files to it at any time,
they would write to the disc and be saved, and the disc was left in an
"open state" so that you could add more files at any time. You were
warned that this type of CD may not be compatable with other CD drives
and if unsure, use the burn once method and be done with it. I have used
this feature before and it works quite well, but I think that would be
good for someone who makes spreadsheets every day and needs to back them
up every day. I do not have much use for a small time file dump available
at all times to do this.

> DVD-RAM is using a DVD just like a floppy diskette or hard disk or USB
> memory stick without any of the frustrations of the inadequacies of
> packet burning.

Thank you for the explanation. I have limited time for computer play and
education, and I easily get fascinated with something new and then spend
way too much time with it. Your explanation helped a lot.

>> but for some reason, the only Nero Linux I can find is version 4 and
>> I don't believe it has the full suite of tools with it.
>
> Why would you not use brasero or k3b for working with CDs and DVDs?

I do now. Brasero is not something I was familiar with until it made it's
appearance in the Linux distros. It is an excellent program and I do now
use it often if I need to burn a disc while in Linux. I haven't used it
for diagnostic tools though and as I can see, it really does not come
with any but sure offers a nice set of disc making tools.

> With DVD-RAM, you just create a UDF file system on the DVD-RAM disc
> and then mount it and use as you would with a file system on a floppy
> diskette, hard disk, or USB memory stick.
>
> <http://ubuntuforums.org/showthread.php?t=1028192>
>
> After mounting the file system, you can, if you are GUI inclined,
> use drag and drop from your favorite file manager.

That is a cool link, thanks. Bookmarked for future usgage. Right now I
have a brand new OS to tame and bend to my will. This will take time but
is very promising. CentOS 6.2 is here and alive!

>> Asus DRW-1814BL (IDE)
>
> So you bought something and do not know what it does?

Dude, with all due respect, do you buy a car and know everything there is
to know about that automobile before you make the purchase? Even how it
would stand up to other vehicles on a Nascar track? I don't. If it does
what I need, looks nice, is affordable, and I like it, then I buy it.
Same goes for optical drives. For all I know this DVD drive might be the
one and only model that can store minute amounts of radioactive
strontium, but since I have no need to store radioactive strontium, I
wouldn't know or really wouldn't care. Before you get offended, bear in
mind that I did begin this paragrahp with "with all due respect" and I
mean that.

> QUOTE
>
> The ASUS DVD Burner enables users to burn any
> DVD+R/-R, DVD+RW/-RW or *DVD-RAM* disc
>
>
> UNQUOTE
>
>> Lite-On DVDRW SOHW-8125 (IDE)
>
> Are you absolutely certain that is 8125 and not 812S?

Who knows. I got Disk Utility in CentOS so let's see what it says. Maybe
it will give me some more information. Oh crap, this version does not
show the optical drives! I do not remember the specific CLI command to
probe disk drives but it is not unfamiliar. Aproopos might help here. I
know there is a command to probe disk drives, I have used it before, but
for now it escapes me. hdparm or something like that? Yes that is it.

Oh you are making me work and this is a new OS and Man, Woman, Wild is a
new episode right now. ...sigh. (I have a TV card in my PC) Crap, no
cdrom symlink in /dev/, great. Wow, hdparm -I give a LOT of information.
Too bad I do not know which device is the optical drive.

Oh crap, the system does not see my opticals at all and one will not even
eject. Dammit. This OS is not even 24 hours old. Loose cable? I have to
go. Will get back to you on this.

> You stated "These burners are within the 2 year mark, I think."
>
> I cannot find any reference to an 8125 and the 812S model
> is from 2004, which is over 7 years old.
>
> <http://www.hardwarereview.net/Reviews/Dual%20Format%20DVD%20Burners/Du
> alFormatBurnerRoundup.htm>
>
> And the ASUS DRW-1814BL is a 2007 model so that makes it 4 years old
> at least, twice the age that you were under the impression it was.
>

Yes, time sure does fly. Probably is more like 4 years when I think about
it. Okay, I got a nice new OS and no optical drives at all. Time to roll
up the sleeves. On the bright side, the nvidia drivers went right in and
the system is quite stable and nice, compiz was right there in desktop
effects and one click enabled all of it, quite nice otherwise. Now where
did these optical drives go? Time to look in the BIOS. Get back to you
soon. Happy New Year JG.

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je32f5$7fu$2...@dont-email.me:

> On Wednesday, January 4th, 2012, at 22:36:09h +0000, Ohmster wrote:
>
>> J G Miller <mil...@yoyo.ORG> wrote in news:je1rsu$d0$2...@dont-email.me:
>>
>>> On Wednesday, January 4th, 2012, at 05:00:55h +0000, Ohmster wrote:

[..]

> QUOTE
>
> The ASUS DVD Burner enables users to burn any
> DVD+R/-R, DVD+RW/-RW or *DVD-RAM* disc
>
>
> UNQUOTE
>
>> Lite-On DVDRW SOHW-8125 (IDE)
>
> Are you absolutely certain that is 8125 and not 812S?

Uh, you busted me on a total newbie mistake that has been in place for
years, JG. The new OS did not "see" my optical drives at all, they just
did not exist. Checking them in the BIOS showed one as "error", no drive
present. Replacing the data cable and reversing the order (Two IDE drives
on one controller) realy messed the machine up, all IDE devices vanished
or became corrupt in the system BIOS. Time to pull drives...

I did not give much though to reversing the data cables as I use cable
select almost exclusively. However, that was originally not the case, I
previously always made one drive a master and one a slave and kept
careful track of which was which and wired them accordingly. The reversed
new data cable indicated that something was seriously awry and these were
new drives, the best you could buy at the time and surprisingly, not that
expensive.

It turns out that the lower Lite-On drive was jumpered as master whilst
the Asus was jupered as cable select. Uh oh, BIG no no. New cable,
jumperd all to cable select and things did clean up quite nicely. As far
as your model numbers go, since I had the drives out, I got you numbers
directly from the devices themselves.

Asus DRW-1814BL
Lite-On DVDRW SOHW8

Previously I gave you information from Disk Utility with misconfigured
jumpers. My bad.

You made me work, you made me miss my show, and you saved my system from
what could have been a very long, maddening problem of CDROM issues.
Thanks JG.

> You stated "These burners are within the 2 year mark, I think."
>
> I cannot find any reference to an 8125 and the 812S model
> is from 2004, which is over 7 years old.
>
> <http://www.hardwarereview.net/Reviews/Dual%20Format%20DVD%20Burners/Du
> alFormatBurnerRoundup.htm>
>
> And the ASUS DRW-1814BL is a 2007 model so that makes it 4 years old
> at least, twice the age that you were under the impression it was.

As far as the age and models go, if this is an argument and you are
determined to win, I concede. This is a new year, a new OS, and I would
like to start it out on a positive note. Did I "pull drives" to check
model numbers before giving you the information? No, I did not come here
for CD issues, other than my Fedora CD would not boot. But, I had CD
issues and your persistance paid off. Maybe the Fedora CD will actually
boot now, not sure but don't care. I love this new CentOS and am dying to
get out and fly it for a while. Oh, and by the way, I for sure have
several cd and dvd symlinks in my /dev/ directory now. You rock dude!

[Ohmster]

Bit Twister <BitTw...@mouse-potato.com> wrote in

news:slrnjg9krr.b...@wb.home.test:

Bit, please do not take this the wrong way. You are *way* more advanced
than I can ever hope to be with Linux systems. That is why I come here
for help and love to hear all you have to say. I probably have more pages
stored from you than anyone else on Usenet. But, like I said, this is not
a mission critical system but losing my stuff would surely suck. I went
over this bridge before by running out of room on my root disk. I was
advised to graft a new hard disk onto /home and cp -a my data over to it.
The trick worked brilliantly with no errors at all so I am familiar with
this particular hardware and setup. I gave all the home space back to the
filesystem and ended up with hundreds of gigs of /home space, how cool is
that? And it was so seemless that you would never be able to tell that it
was on a different disk.

I did not dismiss your rsync lesson, I was overwhelmed by it and blown
away at how much you know about it and how easily you could explain it to
me. As Mr. Spock would say, "Fascinating".

By the time I got your reply, I already had the data backed up, checked
it's integrity, and was satisfied by it. Now I have to build my dream
system back up, config by config, script by script, alias by alias, and
way more things than I can even remember like setled for keeping the
numlock key on. How did I do that again? Oh, it was something in
rc.local, I have a copy of it to go back on.

Thanks to your help and confidence, I am now the proud owner of a brand
new CentOS 6.2 system and I love it. Gosh darn JG Miller caught me in a
total newbie mistake. I had one DVD jumpered to master and one to cable
select. How I got along all this time (years) without noticing this is
amazing. CentOS would not put up with it and gave me NO CDs at all!
Installing a new data cable and switching the drives really messed up the
BIOS, ALL IDE drives were gone! (Two optical drives on one IDE
controller.) Oh man, I could have had some major data corruption because
of that. Talk about luck, knock on wood.

I am tired and distracted by the television, if you have a "brief"
moment, what are acls? I know I can look it up and probably will if you
do not answer but I need to know if I am in danger of data corruption. I
will look it up anyway but after Man, Woman, Wild is over.

Thank you bit twister.

[Ohmster]

unruh <un...@invalid.ca> wrote in news:nw6Nq.2$QF...@newsfe01.iad:

> On 2012-01-04, Ohmster <ro...@dev.nul.invalid> wrote:
>> Bit Twister <BitTw...@mouse-potato.com> wrote in
>> news:slrnjg8nf3.f...@wb.home.test:
>>
>>> On Wed, 4 Jan 2012 08:28:47 +0000 (UTC), Ohmster wrote:
>>>> Bit Twister <BitTw...@mouse-potato.com> wrote in
>>>>>
>> one.
>>
>> I stated my intentions to drop fedora like a hot rock and my plan to
>> backup what I have and put it back to use in a similar, but stable
>> operating system. CentOS looks like the way to go, RHEL was one
>> thought but it is expensive. The idea to pirate RHEL makes me feel
>> like crap, I
>
> Uh, wht is expensive under Redhat is not the release, but the support
> contract. RHEL is a GPL distribution. Ie, it is completely legal for
> anyone to copy it and use it. That is what Linux is all about. It is
> NOT pirating. This is exactly what Centos does. It takes Redhat and
> copies it.

unruh, that is excellent and brief advice. I totally agree with you. Yes,
REHL gives technical support and CentOS does not. It is blatently
pirating that I try to avoid since I have been guilty of it for more PC
games than I can count. I got stopped dead in my tracks with Quake, John
Carmak with his darned CD key validation every time you tried to open the
game while online nearly killed me trying to pirate. After putting nearly
60 hours into trying to pirate it, I realized had I just worked 60 hours
and paid for the game, I could have not only been playing weeks before, I
would have had tons of money left over to pay for stuff with. So much for
pirating.

Since you put it that way, pirating RHEL would be no different than
installing CentOS since I could never call for tech support. I did not
know that CentOS existed and only after compaining relentlesly about
Fedora in here did someone suggest I try CentOS. What a Godsend that was!

>> just cannot do that to the Linux community and then come here and ask
>> for
>
> Do what? Use linux exactly as it is supposed to be used?

Hah, thanks. Open source rules. I often wish that Bill Gates and Steve
Jobs did not capitolize so badly on software. I do not care that they
make money and even became very, very wealthy on software, but because of
the straglehold they have, no one takes open source seriously and this
squashes development of the very best software ever being ported over to
it. Adobe or EA would never give two cents to open source, only because
they have to pay the cost of developement, want their investment back,
and never release their hold on the software even after making their
investment back, several times over. I give credit to Parallax to porting
Decent over to Linux. It played dammed good, every bit as good as
Windows, if not better. I played that game so many times, every version
of it, I had a nightmare where I was suddenly in the ship in the mines
and the missles and enemies were real. That was NOT fun and I woke up
shaking and sweating. Maybe I need to find something else to play for a
while. ;>)

Wow, Descent is now open source and available for linux for free! Supply
your own registered or shareware game files and I have all the registered
game files, bought in advance and paid for. Hmmm, the download links are
all broken and people are having a hard time compiling it.
http://happypenguin.org/show?Descent

>> help. Then I heard about CentOS and my prayers were answered. I
>> looked at the EOL date on Cent 6.2 and see it is 2017! Wow, after
>> getting burned on redhat 9 and letting it go way past EOL and having
>> the script kiddies get in via apache, I desperately want to get out
>> of this update every 6 months stuff. Security updates, fine. Some new
>> features, great, but the entire OS from scratch, every 6 months? Uh
>> no. Fedora is for experimenters and that is pretty cool but I rely on
>> this machine too much for that.
>>
>> Since no one strongly objected to my cp-a, I am almost done with it
>> and am satisfied that I have not lost any of my important stuff. Oh,
>> look at
>
> Well, I would not bet on it. One of the features of rsync is that it
> does a hash of the two files to ensure that they really really are the
> same. It is also much faster and does not copy stuff that has already
> been copied. Start using rsync.

Yes sir. I will take that as an order. I desperatly need a way to backup
my system and backup-manager, as easy as it was to setup, gives almost no
information on how to restore what you have backed up. I have heard
enough about it for a long enough time to know you are right. But, before
I can backup anything, I need something to backup and that means
rebuilding my system over again from saved configs and files. I have a
LOT of work cut out for me so "I'll be back" with lots of rsync questions
when the time arrives.

>> that, it is done! So that is it then. I am going to put the CentOS
>> 6.2 install set in and let 'er rip. The configs should be very close
>> to the same so that should not be a major issue. I know how to work
>> the os and the package system. I am good to go.
>

Thanks unruh.

[Ohmster]

Gordon <Gor...@clear.net.nz> wrote in
news:9ml5gi...@mid.individual.net:

Gordy, I could not have said it better myself. I just got dragged into it
by brand loyalty because I got Redhat 6.0 for Christmas one year, boxed
set, and loved it. First time I could easily install and manage Linux
software and yum was a treat that I cannot live without.

Redhat was true to it's end user, Fedora just drifted away from the end
user and more to the software companies as a testbed to see what would make
them money and what wouldn't. Let the end user be the guinea pig. That
sucks so bad. So much for brand loyalty.

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je1rnt$d0$1...@dont-email.me:

This is awesome JG. I am keeping this and bit twister's tutorial for when
I get my shit back up and right so that I can make a good backup system.
It looks like bit twister has it worked out perfectly but appears overly
complicated to me at first glance. Now HE understands it and can probably
do it blindfolded, but that does not mean that I can understand it or
even accomplish it. Bit Twister is an amazing walking reference manual. I
can probably use his system if I get some shell scripts to acutally run
it, even better if it can be scheduled as a cron job. Thanks buddy.

Where the heck is Steve Ackerman? Is he still pissed at me for being
ignorant or has he dropped out of site? The man is brilliant and very
helpful, but has a short fuse for those that fail to learn quickly. Most
of the really top end tech guys are that way. I am more compassionate but
have been burned pretty badly because of it.

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je2jek$6s7$1...@dont-email.me:

Oh you know it dude. And guess what? My first project was installing
genuine nvidia drivers because I stumled on the perfect tutorial and guess
what is under Desktop Settings since doing that? Compiz with all the bells
and whistles, wobbly windows and desktop cube. One click turned it all on
and it works like a charm! Holy crap! I cannot believe this.

Here is the tutorial if anybody wants it. HowTo install real nvidia drivers
on CentOS 6. What a difference this made!
http://pkgs.org/forum/viewtopic.php?pid=50#p50

[Ohmster]

Bit Twister <BitTw...@mouse-potato.com> wrote in

news:slrnjg7p87.7...@wb.home.test:

Bit, it works! It works like a charm, clean, fast, easy, and sweet. Already
have rean nvidia drivers and this thing is sweet. Proud ownder of a CentOS
6.2 system where all works perfectly!

I have a lot of work to do now but will be back soon. You and I have work
to do on rsync but I need something to backup before I am ready for that.
Your cooker system is cool but you do a LOT of work to run it, is it
possible to make shell scripts and a cron job to automate this backup via
rsync?

(Do not go into great detail now, I am not ready for it, unless you feel
like it. It will not be wasted.)

[Bit Twister]

On Fri, 6 Jan 2012 04:05:36 +0000 (UTC), Ohmster wrote:
> Bit Twister <BitTw...@mouse-potato.com> wrote in


>

> I did not dismiss your rsync lesson, I was overwhelmed by it and blown
> away at how much you know about it and how easily you could explain it to
> me.

It was easy because it is easy for simple partition copies as long as
you remember source has trailing slash and target does not have a
trailing slash.

> Thanks to your help and confidence, I am now the proud owner of a brand
> new CentOS 6.2 system and I love it.

> I am tired

I hear that, just finished a clean install/updates to Mageia Linux Alpha2.

> and distracted by the television,

Heheheh, watching The Big Bang Theory on my pc via MythTv which I had
recorded earlier.

> if you have a "brief" moment, what are acls?

Access Control List. Is that brief enough?

Think of ACLs as file permissions on steroids. You can set permissions
on just about anything and assign those to groups.
You then assign those acl group(s) to users.

Example, you could have a 700 root root /some/directory, and an ACL which
will allow a user to be able to do anything in /some/directory.

Reason I mentioned it, RedHat started pushing it into production and
other distributions are starting to pick it up.

Like systemd, gui tools for managing it, is a bit behind the curve.

As for folding in your changes, I really recommend you put the changes
in scripts. I usually just type the command in a script, highlight/select
from script, paste into a terminal to verify it works.

Next install, I just run the script.

My rc.local is plumb easy, I just link /etc/rc.d/rc.local to my copy.

$ ll /etc/rc.d/rc.local
lrwxrwxrwx 1 root root 19 Aug 31 22:25 /etc/rc.d/rc.local -> /local/bin/rc.local

Same for /etc/profile, /etc/bashrc.

$ ll /etc | grep /local/bin
lrwxrwxrwx 1 root root 17 Aug 31 21:06 bashrc -> /local/bin/bashrc
lrwxrwxrwx 1 root root 18 Aug 31 21:06 profile -> /local/bin/profile

My global environment is via link in /etc/profile.d

$ ll /etc/profile.d/* | grep /local/bin
lrwxrwxrwx 1 root root 22 Aug 31 21:06 /etc/profile.d/xx_local.sh -> /local/bin/xx_local.sh

My login_changes script becomes something like
_fn=/etc/profile
if [ ! -e ${_fn}_vorig ] ; then
mv $_fn ${_fn}_vorig
ln -s /local/bin/profile
fi

_fn=/etc/bashrc
if [ ! -e ${_fn}_vorig ] ; then
mv $_fn ${_fn}_vorig
ln -s /local/bin/bashrc
fi

ln -sf /local/bin/xx_local.sh /etc/profile.d/xx_local.sh

The _vorig is a the vendor original copy, and I test it to see if I
have already made the change.

The _vorig comes in handy to compare a new copy against my mods to see
what changed.

[Bit Twister]

On Fri, 6 Jan 2012 04:48:54 +0000 (UTC), Ohmster wrote:

> Your cooker system is cool but you do a LOT of work to run it,

Only when there are major changes like the systemd change.
Rest of the time is spent checking new config files against my changes
to catch what is new/changed.

> is it possible to make shell scripts and a cron job to automate this
> backup via rsync?

> (Do not go into great detail now, I am not ready for it, unless you feel
> like it. It will not be wasted.)

Ok, no great detail then. :-D

99.9% of the time all your repetitive console commands can be scripted, somehow.

Scripting is like riding a bicycle or sex, you can read all you want about it,
but you have to do it, to be worth a damn. :)

Snippet the following from an up line response you made to

J G Miller <mil...@yoyo.ORG> wrote in news:je1rnt$d0$1...@dont-email.me:

> Now HE understands it and can probably
> do it blindfolded, but that does not mean that I can understand it or
> even accomplish it. Bit Twister is an amazing walking reference manual.

Not really, any command I have to research more than once, I put in
a unix.help text file with some keywords. I have a script which greps
for the supplied key word(s) argument(s). Example:

$ uh rsync
copy current partition to another rsync -aAvx ./ /new
copy one partition to another rsync -aAvx /old/ /new # if do not want /new/old
copy one partition to another rsync -aAvx /src /dest # if want /dest/src
copy one partition to another rsync -aAvx /src/yy/ /dest # if want /dest
list rsync modules rsync ftp.example.edu::
rsync usage rsync ftp.example.edu::modulename
copy partition dir rsync -aAvxz /mnt/cdc5 /mnt/cda2
another backup cmd rsync -raAvx --delete -e ssh /src/ username@new_machine:/dest
copy one partition example mount /alpha ; mount /cooker ; rsync -aAvx --delete /alpha/ /cooker
backup cooker example mount /cooker ; mount /hotbu ; rsync -aAvx --delete /cooker/ /alpha
restore cooker example mount /cooker ; mount /hotbu ; rsync -aAvx --delete /alpha/ /cooker
vbox guest rsync -aAvx --delete --exclude-from=/local/bin/rsync.excludes ro...@wb.home.test:/accounts/ /accounts
vbox guest rsync -aAvx --delete --exclude-from=/local/bin/rsync.excludes ro...@wb.home.test:/local/ /local
vbox guest rsync -aAvx --delete ro...@wb.home.test:/local/doc/ /local/doc
vbox guest rsync -aAvx --delete ro...@wb.home.test:/local/bin/ /local/bin
any_location rsync -aAvx --delete $US...@wb.home.test:$PWD/ $PWD

So, as you can see, I pasted a few commands and told you what they did.
Several of the last ones I use pretty regularly on new installs so I
just copy/paste into a terminal and I am done.

> I
> can probably use his system if I get some shell scripts to acutally run
> it, even better if it can be scheduled as a cron job.

As far as cron maintenance goes, you can make yourself a maintenance
headache or make it easy upon yourself, depending on how you go about it.

I made it simple. I create one link in each of the /etc/cron.*
directories which point to my /local/bin/local_cron_job.

The script gets the app name and tells run-parts to run all the jobs
in that directory.

My cron_changes creates the links and I am done setting it up after a
new install. That assumes /local partition is already on the system.

Anytime I need a new cron job, I just put the script or link to a
script in the required /local/cron/wherever directory I am done with
the change after the next backup of /local.

Some reference info:
/etc/cron.daily has
lrwxrwxrwx 1 root root 25 Aug 31 22:25 _daily -> /local/bin/local_cron_job
/etc/cron.hourly has
lrwxrwxrwx 1 root root 25 Aug 31 22:25 _hourly -> /local/bin/local_cron_job
/etc/cron.monthly has
lrwxrwxrwx 1 root root 25 Aug 31 22:25 _monthly -> /local/bin/local_cron_job
/etc/cron.weekly has
lrwxrwxrwx 1 root root 25 Aug 31 22:25 _weekly -> /local/bin/local_cron_job



$ cat /local/bin/local_cron_job
#!/bin/bash
#******************************************************
#*
#* local_cron_job - cron job to run local/(cron/hourly|daily...) jobs
#*
#*
#* Usually called from /etc/cron.hourly, daily, monthly,.....
#*
#* Usage: link _name /local/bin/local_cron_job
#*
#* Example: cd /etc/cron.hourly
#* ln -s /local/bin/local_cron_job _hourly
#* Leading underscore on _name is important because the underscore
#* causes it to be executed ahead of other scripts and necessary
#* because it is removed to get directory name in /local/cron.
#*
#***********************************************************************

/local/bin/cronlock check
if [ $? -eq 0 ] ; then
exit 0
fi

_exe=$(basename $0)

_job=${_exe:1}

if [ ! -d /local/cron/$_job ] ; then
echo "$0 link is incorrect"
echo "or /local/cron/$_job does not exist"
/bin/false
else
/local/bin/mount_all umount > /dev/null 2>&1
nice -n 19 run-parts --report /local/cron/$_job
/bin/true
fi

#*********************** end local_cron_job *******************************

[Aragorn]

On Friday 06 January 2012 06:09, Bit Twister conveyed the following to
alt.os.linux...

> On Fri, 6 Jan 2012 04:05:36 +0000 (UTC), Ohmster wrote:
>
>> if you have a "brief" moment, what are acls?
>
> Access Control List. Is that brief enough?
>
> Think of ACLs as file permissions on steroids. You can set permissions
> on just about anything and assign those to groups.
> You then assign those acl group(s) to users.
>
> Example, you could have a 700 root root /some/directory, and an ACL
> which will allow a user to be able to do anything in /some/directory.
>
> Reason I mentioned it, RedHat started pushing it into production and
> other distributions are starting to pick it up.

The reason why RedHat is pushing it is because ACLs are a requirement
for SELinux, which RedHat is /also/ pushing.

All readers however please take notice of the fact that just because
RedHat is pushing something, that doesn't necessarily mean that it's the
right thing to do. It does however mean that it is most likely to
become adopted in other distributions, because RedHat carries a lot of
weight. (This in reference to my earlier comments with regard to the
planned move of "/bin/*" to "/usr/bin/*" and "/lib/*" to "/usr/lib/*",
and the concept of making "/usr" into the new "/".)

In this case, ACLs are useful, but not really a requirement for a home-
based system, as there is nothing wrong with the standard UNIX/POSIX
"ugo" permissions model. In absence of ACLs, UNIX always resorts to the
default "ugo" permissions model. ACLs are however also used as the
default security model in openVMS, and by consequence also in the NT-
based versions of Windows, given that the Windows NT kernel is in fact
more or less a copy of the VMS kernel. (Dave Cutler was cought
redhanded copying code from VMS into NT while still at DEC and was fired
on the spot because of it, and subsequently hired by Microsoft as Chief
NT Kernel Architect.)

Bottom line: While ACLs are a necessity in openVMS and Windows (by lack
of any other permissions system there), they are rather a surplus
feature in UNIX, to be used alongside the default UNIX/POSIX
permissions. They are often used to finegrain the permissions on NFS
shares.

[Eef Hartman]

unruh <un...@invalid.ca> wrote:
> Well, I would not bet on it. One of the features of rsync is that it
> does a hash of the two files to ensure that they really really are the
> same.

Only when you use the -c cq --checksum option:
-c, --checksum skip based on checksum, not mod-time & size

Normally files are compared on (last-modified) date-time and size only.

> It is also much faster and does not copy stuff that has already
> been copied. Start using rsync.

With the -c option it DOES get much slower as it has to read all of
the files TO determine those checksums, both in the source and
destination trees.
--
******************************************************************
** Eef Hartman, Delft University of Technology, dept. SSC/ICT **
** e-mail: E.J.M....@tudelft.nl - phone: +31-15-27 82525 **
******************************************************************

[J G Miller]

On Friday, January 6th, 2012, 10:32:54 +0100, Eef Hartman wrote:

> Only when you use the -c cq --checksum option:
> -c, --checksum skip based on checksum, not mod-time & size
>
> Normally files are compared on (last-modified) date-time and size only.

Thanks for pointing that out.

On a couple of occasions, with some small configuration files,
they were not being update on the remote file system because the
size and time was the same on each even though the contents were
slightly different eg something like one had blah = 1024 and
the other had blah = 4096.

Incidentally for those who want to synchronize a lot of directories
and files on a regular basis, with visual feedback and the ability
to skip changes, one should have a look at unison.

<http://www.cis.upenn.EDU/~bcpierce/unison/>

[J G Miller]

On Friday, January 6th, 2012, at 10:00:41h +0100, Aragorn wrote:

> All readers however please take notice of the fact that just because
> RedHat is pushing something, that doesn't necessarily mean that it's the
> right thing to do.

Indeed not, you need Wilfrid Brimley for that!

> In this case, ACLs are useful, but not really a requirement for a home-
> based system, as there is nothing wrong with the standard UNIX/POSIX
> "ugo" permissions model.

If you use individual groups for each user (ie group name = user name)
as has been the case of a number of distributions for quite a long time,
rather than group = users, it is highly unlikely that any home user would
ever need to use ACLs.

As you so well explain, ACLs are needed when you want a very fine
control on a system with lots of users and different groups with
lots of members.

For example, smith in accounting group is allowed access to all the
other accounting group owned files (as per rwx permissions on each)
but for some special project she needs to be able to read certain
files belonging to engineering but not all engineering group readable
files, and this is where ACLs make this possible rather than having
to set up a special group and having to add all of engineering and
smith to that group. And of course when smith leaves that special
project, no need for any group deletion and change of group
ownership, just removal of the ACL for smith.

[Aragorn]

On Friday 06 January 2012 17:35, J G Miller conveyed the following to
alt.os.linux...

> On Friday, January 6th, 2012, at 10:00:41h +0100, Aragorn wrote:
>
>> All readers however please take notice of the fact that just because
>> RedHat is pushing something, that doesn't necessarily mean that it's
>> the right thing to do.
>
> Indeed not, you need Wilfrid Brimley for that!

You had me pull up Wikipedia for that one. ;-)

>> In this case, ACLs are useful, but not really a requirement for a
>> home- based system, as there is nothing wrong with the standard
>> UNIX/POSIX "ugo" permissions model.
>
> If you use individual groups for each user (ie group name = user name)
> as has been the case of a number of distributions for quite a long
> time, rather than group = users, it is highly unlikely that any home
> user would ever need to use ACLs.
>
> As you so well explain, ACLs are needed when you want a very fine
> control on a system with lots of users and different groups with
> lots of members.
>
> For example, smith in accounting group is allowed access to all the
> other accounting group owned files (as per rwx permissions on each)
> but for some special project she needs to be able to read certain
> files belonging to engineering but not all engineering group readable
> files, and this is where ACLs make this possible rather than having
> to set up a special group and having to add all of engineering and
> smith to that group. And of course when smith leaves that special
> project, no need for any group deletion and change of group
> ownership, just removal of the ACL for smith.

Exactly, and that is why UNIX is superior to those operating systems
that solely rely on ACLs. There's always the combination of the "ugo"
permissions mask and the owner/group model to fall back to. ;-)

[J G Miller]

On Friday, January 6th, 2012 at 03:35:39h +0000, Ohmster wrote:

> I did not give much though to reversing the data cables as I
> use cable select almost exclusively.

As you now realize, a really, really bad installation.

For IDE disks and CD/DVDs always use one controller per item and
always as master.

Now that IDE is fading away and SATA is becoming dominant, there
should never be an issue of needing to put more than one device on
an IDE controller because of not enough controllers.

Of course if you have lots of IDE devices, then just buy an
IDE controller card to keep to the rule of one device per
controller.

> As far as the age and models go, if this is an argument and you are
> determined to win, I concede.

It is not a case of trying to win any argument, but getting a set
of consistent facts. Without consistent details, trying to offer
any help becomes frustrating and pointless.

Soooooooo it is in your own best interests to give
accurate information. (Sorry if that sounds so schoolteacherish.)

> Oh, and by the way, I for sure have several cd and dvd symlinks
> in my /dev/ directory now.

Just a note that on older kernel versions it used to be the case
that scd<dev_number> was sometimes used as in intermediate link
to refer to CD/DVD devices under the libsata scheme but now under
Linux 3, sr<dev_number> seems to be the only thing used.

Also access to mounted CD/DVD devices tends to be under the
/media/dvd<whatever> mount point (often autmoagically mounted)
rather than /mnt.

Hope you get your system up fully customized, and I am sure
that with CentOS you will have a stable and far less frustrating
time than bleeding edge Fedora.

Also it is worth mentioning that bleeding edge distributions tend
to be geared towards exploiting the latest hardware, dropping
support for older hardware items, so again if you are not running
it on the latest machine, it is not such a good idea.

[J G Miller]

On Friday, January 6th, 2012 at 02:52:14h +0000, Ohmster wrote:

> Dude, with all due respect, do you buy a car and know everything there
> is to know about that automobile before you make the purchase?

Whenever I buy anything computer/electronic, I draw up a mental
or written list of all the features that I need, nice to have,
and latest technology features that may be useful in the future
and then do a comparison of models, preferably consulting some
reviews at a reputable site. For CD/DVD drives this would be
www.cdrinfo.COM and perhaps a check on Tom's Hardware and
Afterdawn.com also.

I have been caught out a couple of times before because of not
being informed about the product before buying because it looked
like it was "too good a price to miss" and it appeared to do
the job I needed it

> If it does what I need, looks nice, is affordable, and I like it,
> then I buy it. Same goes for optical drives.

So you admit you are an uninformed consumer, which is the exactly
the sort of consumer that the sales and marketing people rely on
to make their sales and profits.

Remember "Caveat Emptor".

[J G Miller]

On Friday, January 6th, 2012, at 04:33:09h +0000, Ohmster wrote:

> Redhat was true to it's end user, Fedora just drifted away from the end
> user and more to the software companies as a testbed to see what would
> make them money and what wouldn't.

Which software companies would those be?

Fedora is much more stringent in excluding proprietary software
(even MP3 decoding because of the Fraunhoffer licensing issue)
from its distribution than many other distribution.

[unruh]

Well, to be fair, acls are a much more generic permission model than the
usual unix model. You can give various groups and individuals
independent access permissions. Thus you can make the file unuseable
for everyone, readable for Joe, Mike and Tom, executable and readable
for Peter and Sam , also writeable for yourself and Mary, but finally
owned by root, and suid root.

That is impossible under the usual unix permissions. Now, for most
files, it does not really matter, but it is that small subset of cases
where it can be really useful. Under VMS/NT, this is the default. There
never was a call for the unix model, because they instituted this more
generic model by default. It was not "they had nothing so grafted on
acls" but "they designed the system to use acls from the start". Of
course because of the history of windows, where nothing was the default
for so long, most users kept on that way, and ignored ( or rather did
not even know that they had them or what to do with them) acls and
assumed an "open"model. And MS did not enforce the ACLs model in
installation because of "backwards incompatibility".
Fine for an individual machine, not so fine for
a machine connected to the internet.


>

[unruh]

On 2012-01-06, Eef Hartman <E.J.M....@tudelft.nl> wrote:
> unruh <un...@invalid.ca> wrote:
>> Well, I would not bet on it. One of the features of rsync is that it
>> does a hash of the two files to ensure that they really really are the
>> same.
>
> Only when you use the -c cq --checksum option:
> -c, --checksum skip based on checksum, not mod-time & size

There are two questions-- does rsync start the transfer of material, and
is the material transfered the same as what was to be transfered. rsync
always does the latter. rsync uses the checksum to decide whether or not
to start a transfer as you say only if y ou use the -c option. Thus,
rsync, in default mode, could decide not to transfer file A from M1 to
M2 because the size and date of A on M1 and M2 are the same, but the two
files could be different. But once it has decided to transfer, then it
uses the checksum both to decide how much to transfer ( onely those
parts of the file that are different)and to make sure that the final
result is really the same on M1 as on M2

>
> Normally files are compared on (last-modified) date-time and size only.
>
>> It is also much faster and does not copy stuff that has already
>> been copied. Start using rsync.
>
> With the -c option it DOES get much slower as it has to read all of
> the files TO determine those checksums, both in the source and
> destination trees.

Yes. So people usually decide that the size and date are sufficient to
decide whether or not to transfer. But once the decision to transfer has
been made, the files have to be read anyway, and the checksumming takes
no more time ( less, since only those parts of the file that have
different checksums need to be transfered, and it is transfering that
takes the time, not checksumming-- at least that is the rsync model.)

[unruh]

On 2012-01-06, J G Miller <mil...@yoyo.ORG> wrote:
> On Friday, January 6th, 2012 at 03:35:39h +0000, Ohmster wrote:
>
>> I did not give much though to reversing the data cables as I
>> use cable select almost exclusively.
>
> As you now realize, a really, really bad installation.
>
> For IDE disks and CD/DVDs always use one controller per item and
> always as master.
>
> Now that IDE is fading away and SATA is becoming dominant, there
> should never be an issue of needing to put more than one device on
> an IDE controller because of not enough controllers.
>
> Of course if you have lots of IDE devices, then just buy an
> IDE controller card to keep to the rule of one device per
> controller.

Except most motherboards do not have enough slots for more than about 1
ide controller. So you would also have to replace your motherboard.
Cheaper to buy a new SATA disk.

[unruh]

On 2012-01-06, J G Miller <mil...@yoyo.ORG> wrote:
> On Friday, January 6th, 2012 at 02:52:14h +0000, Ohmster wrote:
>
>> Dude, with all due respect, do you buy a car and know everything there
>> is to know about that automobile before you make the purchase?
>
> Whenever I buy anything computer/electronic, I draw up a mental
> or written list of all the features that I need, nice to have,
> and latest technology features that may be useful in the future

For computer equipment, never buy on the basis of the future. The future
will change in ways you had not anticipated and the price will come
down. Buy on the basis of what you need now.

> and then do a comparison of models, preferably consulting some
> reviews at a reputable site. For CD/DVD drives this would be
> www.cdrinfo.COM and perhaps a check on Tom's Hardware and
> Afterdawn.com also.
>
> I have been caught out a couple of times before because of not
> being informed about the product before buying because it looked
> like it was "too good a price to miss" and it appeared to do
> the job I needed it
>
>> If it does what I need, looks nice, is affordable, and I like it,
>> then I buy it. Same goes for optical drives.

The question is, how do you answer "If it does what I need"? How do you
determine that? It is rarely possible to take it for a test drive in the
case of computer equipment. Would you buy a car where the salesman
refused to allow you to take it for a test drive?

[unruh]

On 2012-01-06, J G Miller <mil...@yoyo.ORG> wrote:

> On Friday, January 6th, 2012, at 04:33:09h +0000, Ohmster wrote:
>
>> Redhat was true to it's end user, Fedora just drifted away from the end
>> user and more to the software companies as a testbed to see what would
>> make them money and what wouldn't.
>
> Which software companies would those be?

Redhat.
Fedora is Redhat's testbed.

>
> Fedora is much more stringent in excluding proprietary software
> (even MP3 decoding because of the Fraunhoffer licensing issue)
> from its distribution than many other distribution.

Is this supposed to be an example of them drifting further away from the
user? Sounds like it to me.
It is like the stupidity of not having cdrtools on their distro.

>

[J G Miller]

On Friday, January 6th, 2012, at 17:22:23h +0000, Unruh wrote:

> Except most motherboards do not have enough slots for more than about 1
> ide controller.

In the past motherboards generally had 2 IDE controllers.

Installing a single IDE PCI card would then provide another 2 controllers.

Most people would not have more than 4 IDE devices in the PC,
eg 2 disks and 2 CD/DVDs.

> Cheaper to buy a new SATA disk.

Well as I hinted, in the post IDE world where most motherboards
have at least 4 independent SATA connections, and SATA disks are
generally cheaper per Gbyte than IDE disks if those are still even
on the shelf, there is no point in purchasing an IDE disk.

And CD/DVD burners increasingly have SATA rather than IDE
interfaces.

[J G Miller]

On Friday, January 6th, 2012, at 18:08:45h +0000, Unruh wrote:

> For computer equipment, never buy on the basis of the future. The future
> will change in ways you had not anticipated and the price will come
> down. Buy on the basis of what you need now.

Which is why I listed that point as the third and least important.

But the point is that with some things in the future, it is very clear
what is coming eg in Europe do not buy a new TV without a DVB-t2 tuner
if you do not want an obsolete terrestrial received even if DVB-t2
transmissions are not yet available in your area.

As for computer components, if the feature is available now, then
why would you not want to buy the model which has it, provided that
the other criteria are met and the price is not significantly different?

> The question is, how do you answer "If it does what I need"?

Read reviews and learn from the experience of other people.

[unruh]

On 2012-01-06, J G Miller <mil...@yoyo.ORG> wrote:

> On Friday, January 6th, 2012, at 18:08:45h +0000, Unruh wrote:
>
>> For computer equipment, never buy on the basis of the future. The future
>> will change in ways you had not anticipated and the price will come
>> down. Buy on the basis of what you need now.
>
> Which is why I listed that point as the third and least important.
>
> But the point is that with some things in the future, it is very clear
> what is coming eg in Europe do not buy a new TV without a DVB-t2 tuner
> if you do not want an obsolete terrestrial received even if DVB-t2
> transmissions are not yet available in your area.

Not sure what DVB-t2 is, but here in Canada I bought a dvd TV recorder
about 2 years before the transition to digital tv, and loh, the tuner
was purely analog. 6 months after the transition I went into one of our
big electronics retailers and asked about DVD tv recorders, and they
were still all analog tuners, even though analog transmission had ceased
6 months earlier. So, yes, if standards are going to change in the
immediate future, it is probably good to get something that will comply
with those standards.
On the other hand buying a blueray dvd 5 years ago because bluray was
going to be the next big thing was probably a very bad idea.

also, TVs tend to be a pretty mature market, like cars were. So buying
on future plans can be OK. It tends to be a bad idea in rapidly evolving
areas like most computational stuff.

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je79oc$8pc$4...@dont-email.me:

ROTFLMAO!

Oh God, I am in tears really laughing at this one, you caught me, I have
to wear it, but will bite and wear it with whatever honor there is with
such an act, admission perhaps?

You know, I kind of forgot about the common Internet acronym used above.
I happened to be watching and episode of Law & Order, Special Victims
Unit, Season 6, Episode 14, "Game". It was a really good episode for me
because it involved PC gaming. The victim was run over by a car, stomped
to death with spiked heels, purse robbed, and what were detectives
Stabler and Benson to do? It turns out that Stabler's kid son is a fan of
video game and was visiting daddy at work in the squad room and saw some
of the crime scene photos on the wall. The kid asked about it, the photos
were grisly, and dad briefly explained it and covered the photos to
protect his kid. Somehow the kid noticed the girl was run over, kicked to
death, and was robbed of her purse and he immediately says this is the
hottest new game on the Internet, "Intensity". He then shows the game to
dad and plays the game, viciously running down an helpless woman and
gleefully showing dad how moneybags appear as points, he says he must get
the purse for bonus points and grabs it. More moneybags appear and dad is
totally shocked.

To make a long story short, they find the perps who are kids that are
fanatical video PC gamers and their defense is that "The game made them
do it". What is interesting is that the detectives find a programmer that
worked on the game and was fired for putting "bunny porn" in a kiddie
game. He was kept on as a tester but was a brilliant programmer. He was
psyched out by these kids that discuss the game online to put more Easter
Eggs in the game and they pay him like $1,000 to insert certain scenes
into the game. Once caught, the programmer is really sweating, he has a
powerful TV Degaussing coil by the door that "erases his entire hard
disk" when the detectives remove the PC and walk out his door with it.
The perpetrator kids taunt the programmer so bad as to commit suicide by
emailing him faked scenes from the game with his face in it and
"ROTFLMAO" in huge letters on his PC screen as he sees this message.

Okay, not totally related, but it explains my humor and the ancient
acronym used, plus gives you enough information to hunt the episode down
and watch it, if you are so inclined. All episodes available legally here
for free in reasonable quality without horrible ads right here:
http://watchsvu.net/

Dam you handled that quite well, Mr. Miller. I will admit that I was
slightly annoyed with your incessant request for details regarding my
optical drives. Why in hell is this man so concerned about DVD drives in
my system when I came here for distro advice and support? Well, I did
mention I could not boot a seemingly "good", bootable cd in this machine,
could that be it? Who knows, but I came here for advice and one thing the
Linux newsgroups are known for is that they are hungry for details and
will get irritated and not helpful very quickly if not provided with
requested system details. Okay, I'll bite.

I now understand your need for details and irritation that I could not
provide them at will. You make good sense, I got chewed out and it was on
target, but it was not done maliciously but in a well intentioned lesson.
And, as a bonus, you helped uncover a very, very serious DVD
configuration jumper error that could have caused radical data corruption
on all of my IDE devices. One drive cable select and one drive master.
Ouch! Installing a new data cable and not contorting it to match the
drive order really freaked my system out. In the BIOS I had NO IDE drives
at all or strange binary characters for ALL IDE devices. Plus, as per
your insistence for details, I had NO optical drives available on my
system and NO symlinks that deal with optical drives, at all, in my /dev
directory. What a great lesson! I pulled them, I set all to cable select
and all features and functions returned. No "serious" harm done.

What is puzzling me is that all of my configs are good and usable, for
some reason, my original .vimrc file is corrupted. I would love to know
how or why this occurred, it is almost like it was saved as a PC file and
not a UNIX file, but I cannot find any sort of text viewer, Windows or
Linux, that can read it properly. It is all there but there are strange
marks at the end of lines such as ^M and one other mark, some editors )
Notepad ++) show it as the word "sum" in a box, another as a subscript
letter T. The file command IDs it as a data file. Strange. This is not
serious, vim works just fine, but I did like the fact that my mouse used
to work with it, I had aspell installed, and boxes as well, although I
never really use boxes. That was at the infamous "Al Conner's" request.
Strange bird. Back stabber too. Puzzling character.

I just got apache working and am pleased, out of the blue I have a web
customer that needs site changes and oh I am soooo not ready for this. He
is nice, paid up front, and patient. Wonderful client. I need to config
and activate samba, apache (done), and, that will do for the moment. Swat
and Webmin make samba shares really easy, I can use that or maybe just
use my old samba.conf file if it is alright. I do not copy them over, I
run two putty windows side by side to manually copy and paste pertinent
info into the config files so as not to miss some subtle change in the
file. So far they are almost 100% identical between Fedora 15 and CentOS
6.2. Sweet.

I do not have time to investigate this right now (Customer paid for
work.) but if you are so inclined to view this strange .vimrc file, here
it is, original file, renamed vimrc.txt to make it accessible on a web
server. I cannot tell if this file was opened and saved in a non-UNIX
format or if it is genuine data corruption.
http://www.home.comcast.net/~theohmster/Linux/vimrc.txt

Hey thanks JG, I have to go now. Thanks a lot for helping me understand
the value of researching things more thoroughly before making a purchase.
(I am not totally blind, the Lite-On drives have great reviews in Maximum
PC magazine and that rag is pretty darned good for PC and gamer
enthusiasts. They even created a "Maximum Linux" magazine, I subscribed
right away, but they had to drop it, no advertisers would support it.
Perhaps because much of Linux is open source? Hard to sell software in a
rag like that, I guess.)

Forcing me out of my armchair and rolling up my sleeves to pull drives
was a huge plus, that config error was quite serious. Thank you and
goodnight.

[Ohmster]

unruh <un...@invalid.ca> wrote in news:NMGNq.25534$kp3.13454
@newsfe13.iad:

> On 2012-01-06, J G Miller <mil...@yoyo.ORG> wrote:
>> On Friday, January 6th, 2012 at 02:52:14h +0000, Ohmster wrote:

[..]

>> Whenever I buy anything computer/electronic, I draw up a mental
>> or written list of all the features that I need, nice to have,
>> and latest technology features that may be useful in the future
>
> For computer equipment, never buy on the basis of the future. The
future
> will change in ways you had not anticipated and the price will come
> down. Buy on the basis of what you need now.

Could not agree more. Good tip.

[..]

>>
>>> If it does what I need, looks nice, is affordable, and I like it,
>>> then I buy it. Same goes for optical drives.
>
> The question is, how do you answer "If it does what I need"? How do you
> determine that? It is rarely possible to take it for a test drive in
the
> case of computer equipment. Would you buy a car where the salesman
> refused to allow you to take it for a test drive?

Of course not but manufacturers, at least well established ones, rarely
blatantly lie on unit specifications and that is what I base my sales
information on, or perhaps a well respected magazine review or even good
advice from people I know and respect. You, Bit Twister, JG, aho, and
many other fine people I have had the pleasure to meet here.

Notice that I said "well established". I once bought a car stereo
amplifier that advertised 200 watts and it was fairly inexpensive. Since
I bought this device at lunch while at work (I am an electronic
techician.), I immediately hooked it up and tried it out, right on the
bench. Kind of small amplifier, size of a double pack of cigarettes, and
no apparent heat sinks. How could they do this? This "amp" was not loud,
AT ALL. I bench tested it with load and measuring equipment. Seven watts
RMS at eight ohms. My God, am I reading this right?? Check the box, "200
watts" but the measuring spec as not listed. Hmmm...

I opened the unit to inspect inside. Two tiny ICs on a small PC board. I
looked up the ICs. Single channel audio amplifer, rated power 7 watts RMS
at 8 ohms. Wait a minute, how could they just out and out lie like
this?!! There are many ways to measure electical and audio power, some
are not really relevant but can really pump up the numbers. In audio, the
only real spec that means anything to an audiophile is root mean square
or RMS at stated ohm load. One speaker is typically 8 ohms, could be
different. Two 4 ohm speakrs in paralell would really present a 2 ohm
load. Measuing the sine wace as a peak to peak value and then calculating
the power for a 2 ohm load, did indeed equal 200 watts, but this was a
total fake out and they really should be ashamed for stating such specs
as being meaningful or accurate. I returned the device right after work
for a rull refund and the salesman was not surprised, he gave me my money
back and showed me some "real" power amps, too bad I could not afford one
at the time. Yes they had heat sinks and were larger and heavier. No
surpise now. But I was green back then, educated enough to not believe
the specs, but young enough to dream it could be true. Moral, do not
believe all you read and test it yourself, OR, purchase from a well
respected manufacturer or vendor.

Gotta go, nice talking to you, unruh. Cheers buddy. :)

[Ohmster]

J G Miller <mil...@yoyo.ORG> wrote in news:je7fdk$b4v$1...@dont-email.me:

> On Friday, January 6th, 2012, at 18:08:45h +0000, Unruh wrote:
>

[..]

Guys, as you know, I cannot comment on every post, but I do read them and
thank you all very much for your input and discussion. Besides, one web
forum member politely called me "loquacious" and I was at first insulted
but had to look that up in a dictionary before replying. The guy was right,
it is true. I just talk to dammed much.


They hate having to read an encyclopedia just to offer some help. I cannot
blame them, who would?

;>)

[Dan C]

On Sat, 07 Jan 2012 05:19:49 +0000, Ohmster wrote:

> I just talk to dammed much.

Not to mention the innane babbling.


--
"Ubuntu" -- an African word, meaning "Slackware is too hard for me".
"Bother!" said Pooh, as his rectum exploded.
Usenet Improvement Project: http://twovoyagers.com/improve-usenet.org/
Thanks, Obama: http://brandybuck.site40.net/pics/politica/thanks.jpg

[Aragorn]

On Saturday 07 January 2012 07:05, Dan C conveyed the following to
alt.os.linux...

> On Sat, 07 Jan 2012 05:19:49 +0000, Ohmster wrote:
>
>> I just talk to dammed much.
>
> Not to mention the innane babbling.

I'll take Ohmster's inane babbling over that of "T i m" any day, though.
;-)

At least Ohmster uses GNU/Linux, is both friendly and grateful, and
follows up on the advice you give him, even doing his homework in the
process. That other guy on the other hand... <grin>

[J G Miller]

On Saturday, January 7th, 2012 at 04:54:13h +0000, Ohmster explained:

> It is all there but there are strange marks at the end of lines
> such as ^M

Hopefully your .vimrc is not otherwise corrupted and the ^M simply
occurs because the file was edited and saved under windoze.

There are various ways to fix the problem but probably the
safest is just to use the dos2unix utility to convert it from
its current Windoze file state to standard Unix/Linux file state.

In fact when you try to edit it with vim, it should indicate at
the bottom that it is in DOS format.

Windoze insists on having ^M (CTRL-M) characters at the end of each
line and often no new line character at the end of the file.

[Dan C]

Well, that's all true, I guess.

Sure seems like it takes "Ohmster" forever and a day to learn anything,
though. He's been told that Fedora isn't right for him for *YEARS*, but
just doesn't do anything about it (until now, apparently). I sure don't
understand the blind obsession with RH/RPM stuff though. There's so much
better stuff available. <shrug>

--
"Ubuntu" -- an African word, meaning "Slackware is too hard for me".

"Bother!" said Pooh, as he stepped into the acceleration chamber.

[Aragorn]

On Saturday 07 January 2012 23:49, J G Miller conveyed the following to
alt.os.linux...

But... Windoze also insists on an end-of-file character (^Z) _in_ the
file...

You can always trust Microsoft to take something logical and then mess
it up so badly that even its mother wouldn't recognize it anymore.

[Jasen Betts]

Usually the the end-of-file marker is optional, only a in a few rare
cases was it needed.

It's a hold over from CP/M (and possibly DOS version 1(?)) which only
counted file size in allocation units I don't recall if a CP/M text
file that exactly filled a number of clusters needed a cluster full of
^Z to terminate it or not.

--
⚂⚃ 100% natural

--- Posted via news://freenews.netfront.net/ - Complaints to ne...@netfront.net ---

[Aragorn]

On Sunday 08 January 2012 11:03, Jasen Betts conveyed the following to

alt.os.linux...

> On 2012-01-08, Aragorn <str...@telenet.be.invalid> wrote:
>
>> On Saturday 07 January 2012 23:49, J G Miller conveyed the following
>> to alt.os.linux...
>>
>>> On Saturday, January 7th, 2012 at 04:54:13h +0000, Ohmster
>>> explained:
>>>
>>>> It is all there but there are strange marks at the end of lines
>>>> such as ^M
>>>
>>> Hopefully your .vimrc is not otherwise corrupted and the ^M simply
>>> occurs because the file was edited and saved under windoze.
>>>
>>> There are various ways to fix the problem but probably the
>>> safest is just to use the dos2unix utility to convert it from
>>> its current Windoze file state to standard Unix/Linux file state.
>>>
>>> In fact when you try to edit it with vim, it should indicate at
>>> the bottom that it is in DOS format.
>>>
>>> Windoze insists on having ^M (CTRL-M) characters at the end of each
>>> line and often no new line character at the end of the file.
>>
>> But... Windoze also insists on an end-of-file character (^Z) _in_
>> the file...
>>
>> You can always trust Microsoft to take something logical and then
>> mess it up so badly that even its mother wouldn't recognize it
>> anymore.
>
> Usually the the end-of-file marker is optional, only a in a few rare
> cases was it needed.

It was inconsistent, yes. And it was certainly needed for plain text
files, but it didn't always show up in the editor, even if it was there.
Yet another inconsistency.

> It's a hold over from CP/M (and possibly DOS version 1(?)) which only
> counted file size in allocation units I don't recall if a CP/M text
> file that exactly filled a number of clusters needed a cluster full of
> ^Z to terminate it or not.

Yes, DOS was itself a not-too-legal 16-bit "rewrite" of CP/M - which
itself also had the ^Z end-of-file marker - by Tim Paterson of Seattle
Computer, who called it QDOS ("Quick & Dirty Operating System"). When
Willy Gates was approached by IBM for the second time with regard to
providing an operating system [*] for the IBM Personal Computer, he
bought QDOS from Paterson and hired him as Chief MS-DOS Developer.

CP/M itself was loosely inspired by both UNIX and VMS. And NT is a
virtually identical copy of the VMS kernel, and still carries a lot of
the DOS legacy in it, both from DOS directly and from OS/2, which was
devised - by IBM, not by Microsoft - to be the successor of DOS for
80286 and later processors. Here too, Gates pulled the same stunt, i.e.
he approached Dave Cutler, one of the authors of the VMS kernel, to
write NT. Cutler copied "his" VMS code almost verbatim in NT, and was
caught in the act by DEC, who fired him on the spot and sued Microsoft.
Gates settled the matter with DEC out of court and immediately hired
Cutler as the Chief NT Architect.

The Wikipedia article on the subject does of course not mention the
copyright infringement by Cutler - he co-wrote the VMS kernel, but VMS
was the intellectual property of DEC, not of Cutler - or the possible
patent infringement against Digital Research by Paterson, but as we
know, everyone can create and edit Wikipedia pages, and so those are
often full of propaganda. ;-)

[Aragorn]

On Sunday 08 January 2012 18:54, Aragorn conveyed the following to

Forgot the footnote, so here it is... :p

[*] Insofar as either CP/M or DOS could be considered operating systems,
because they operated in real mode, and they were essentially only
application loaders. Once an executable was loaded into memory,
both DOS and CP/M simply pass control over the entire machine on to
that application, making the application itself the operating system
until it exited again.

[Richard Kettlewell]

Aragorn <str...@telenet.be.invalid> writes:
> On Sunday 08 January 2012 11:03, Jasen Betts conveyed the following to
> alt.os.linux...
>> On 2012-01-08, Aragorn <str...@telenet.be.invalid> wrote:

>>> But... Windoze also insists on an end-of-file character (^Z) _in_
>>> the file...
>>>
>>> You can always trust Microsoft to take something logical and then
>>> mess it up so badly that even its mother wouldn't recognize it
>>> anymore.
>>
>> Usually the the end-of-file marker is optional, only a in a few rare
>> cases was it needed.
>
> It was inconsistent, yes. And it was certainly needed for plain text
> files, but it didn't always show up in the editor, even if it was there.
> Yet another inconsistency.

The notion that Windows "insists" on a final ^Z is certainly nonsense,
including for plain text files.

> CP/M itself was loosely inspired by both UNIX and VMS. And NT is a
> virtually identical copy of the VMS kernel, and still carries a lot of
> the DOS legacy in it, both from DOS directly and from OS/2, which was
> devised - by IBM, not by Microsoft - to be the successor of DOS for
> 80286 and later processors. Here too, Gates pulled the same stunt, i.e.
> he approached Dave Cutler, one of the authors of the VMS kernel, to
> write NT. Cutler copied "his" VMS code almost verbatim in NT, and was
> caught in the act by DEC, who fired him on the spot and sued Microsoft.
> Gates settled the matter with DEC out of court and immediately hired
> Cutler as the Chief NT Architect.

And your evidence for these claims is...

--
http://www.greenend.org.uk/rjk/

[J G Miller]

On Sunday, January 8th, 2012, at 18:54:41h +0100, Aragorn alleged:

> And NT is a virtually identical copy of the VMS kernel

Have you seen and compared the source code for each, or do
you have a reputable source for this statement?

[Aragorn]

On Sunday 08 January 2012 19:23, Richard Kettlewell conveyed the

following to alt.os.linux...

> Aragorn <str...@telenet.be.invalid> writes:
>> On Sunday 08 January 2012 11:03, Jasen Betts conveyed the following
>> to alt.os.linux...
>>> On 2012-01-08, Aragorn <str...@telenet.be.invalid> wrote:
>
>>>> But... Windoze also insists on an end-of-file character (^Z) _in_
>>>> the file...
>>>>
>>>> You can always trust Microsoft to take something logical and then
>>>> mess it up so badly that even its mother wouldn't recognize it
>>>> anymore.
>>>
>>> Usually the the end-of-file marker is optional, only a in a few rare
>>> cases was it needed.
>>
>> It was inconsistent, yes. And it was certainly needed for plain text
>> files, but it didn't always show up in the editor, even if it was
>> there. Yet another inconsistency.
>
> The notion that Windows "insists" on a final ^Z is certainly nonsense,
> including for plain text files.

Really? Like I said, just because it doesn't show _in_ the editor
doesn't mean that it isn't there.

>> CP/M itself was loosely inspired by both UNIX and VMS. And NT is a
>> virtually identical copy of the VMS kernel, and still carries a lot
>> of the DOS legacy in it, both from DOS directly and from OS/2, which
>> was devised - by IBM, not by Microsoft - to be the successor of DOS
>> for 80286 and later processors. Here too, Gates pulled the same
>> stunt, i.e. he approached Dave Cutler, one of the authors of the VMS
>> kernel, to write NT. Cutler copied "his" VMS code almost verbatim in
>> NT, and was caught in the act by DEC, who fired him on the spot and
>> sued Microsoft. Gates settled the matter with DEC out of court and
>> immediately hired Cutler as the Chief NT Architect.
>
> And your evidence for these claims is...

I'm not the only one who knows these things. Besides, there are records
of DEC having sued Microsoft over this issue. Just because the case was
dropped again due to Bill Gates's motion to settle matters out of court
- which he did by both paying a large sum of money to DEC and by
promising to support DEC's Alpha processor in Windows NT - doesn't mean
that the records of the complaints were stricken.

[Aragorn]

On Sunday 08 January 2012 20:22, J G Miller conveyed the following to
alt.os.linux...

I have not personally seen the source code, but I do personally know of
people who have. The names of function calls et al were changed for the
NT kernel, but other than that it's roughly the same code.

[Aragorn]

On Sunday 08 January 2012 20:28, Aragorn conveyed the following to

alt.os.linux...

> On Sunday 08 January 2012 20:22, J G Miller conveyed the following to
> alt.os.linux...
>
>> On Sunday, January 8th, 2012, at 18:54:41h +0100, Aragorn alleged:
>>
>>> And NT is a virtually identical copy of the VMS kernel
>>
>> Have you seen and compared the source code for each, or do
>> you have a reputable source for this statement?
>
> I have not personally seen the source code, but I do personally know
> of people who have.

s/of/from

[Richard Kettlewell]

Aragorn <str...@telenet.be.invalid> writes:
> Richard Kettlewell conveyed the following to alt.os.linux...

>> The notion that Windows "insists" on a final ^Z is certainly nonsense,
>> including for plain text files.
>
> Really? Like I said, just because it doesn't show _in_ the editor
> doesn't mean that it isn't there.

Hexdump of a file created in Notepad:

00000000 62 6c 61 68 20 62 6c 61 68 20 62 6c 61 68 0d 0a |blah blah blah..|
00000010 62 6c 61 68 20 62 6c 61 68 0d 0a |blah blah..|

Note the absence of a trailing 1a.

Where did you get the idea that the trailing ^Z was mandatory from?

>>> CP/M itself was loosely inspired by both UNIX and VMS. And NT is a
>>> virtually identical copy of the VMS kernel, and still carries a lot
>>> of the DOS legacy in it, both from DOS directly and from OS/2, which
>>> was devised - by IBM, not by Microsoft - to be the successor of DOS
>>> for 80286 and later processors. Here too, Gates pulled the same
>>> stunt, i.e. he approached Dave Cutler, one of the authors of the VMS
>>> kernel, to write NT. Cutler copied "his" VMS code almost verbatim in
>>> NT, and was caught in the act by DEC, who fired him on the spot and
>>> sued Microsoft. Gates settled the matter with DEC out of court and
>>> immediately hired Cutler as the Chief NT Architect.
>>
>> And your evidence for these claims is...
>
> I'm not the only one who knows these things. Besides, there are
> records of DEC having sued Microsoft over this issue. Just because
> the case was dropped again due to Bill Gates's motion to settle
> matters out of court - which he did by both paying a large sum of
> money to DEC and by promising to support DEC's Alpha processor in
> Windows NT - doesn't mean that the records of the complaints were
> stricken.

I asked for evidence, not elaborated rumour. e.g. who 'knows' them and
how, when was the court case, etc.

That there was design influence is uncontroversial, actual copying of
code is another matter entirely (and seems pretty implausible on purely
pragmatic grounds, actually).

--
http://www.greenend.org.uk/rjk/

[VWWall]

My first "real" computer was a "luggable" Osborne I. It had two single
density 5 1/4 inch floppy disks , (92k each), and ran CP/M. It came
with some useful applications like Word Star and SuperCalc. The CPU was
a Z80, quite advanced for those days!

I have a hard copy of CP/M 2.2. It's not in digital format, and being a
late generation of analog copying, a scanner doesn't work very well with
it. When the first DOS releases appeared, I spent some time comparing
it with DOS. There are some interesting similarities. For a time, IBM
offered CP/M as an alternate to DOD with the PC. It was priced at about
ten times DOS and never sold well.

I have a "boxed set" of OS2, another IBM flop, which I've never installed.

--
Virg Wall

[Aragorn]

On Sunday 08 January 2012 21:01, Richard Kettlewell conveyed the

following to alt.os.linux...

> Aragorn <str...@telenet.be.invalid> writes:
>> Richard Kettlewell conveyed the following to alt.os.linux...
>
>>> The notion that Windows "insists" on a final ^Z is certainly
>>> nonsense, including for plain text files.
>>
>> Really? Like I said, just because it doesn't show _in_ the editor
>> doesn't mean that it isn't there.
>
> Hexdump of a file created in Notepad:
>
> 00000000 62 6c 61 68 20 62 6c 61 68 20 62 6c 61 68 0d 0a |blah
> blah blah..|
> 00000010 62 6c 61 68 20 62 6c 61 68 0d 0a |blah
> blah..|
>
> Note the absence of a trailing 1a.
>
> Where did you get the idea that the trailing ^Z was mandatory from?

I didn't exactly say that it was mandatory. I said that it was needed,
for whatever value of "needed". And I was rather speaking of DOS than
of Windows. Could be that it was just a requirement of the editors in
DOS - EDLIN.COM, EDIT.COM and friends.

>>>> CP/M itself was loosely inspired by both UNIX and VMS. And NT is a
>>>> virtually identical copy of the VMS kernel, and still carries a lot
>>>> of the DOS legacy in it, both from DOS directly and from OS/2,
>>>> which was devised - by IBM, not by Microsoft - to be the successor
>>>> of DOS for 80286 and later processors. Here too, Gates pulled the
>>>> same stunt, i.e. he approached Dave Cutler, one of the authors of
>>>> the VMS kernel, to write NT. Cutler copied "his" VMS code almost
>>>> verbatim in NT, and was caught in the act by DEC, who fired him on
>>>> the spot and sued Microsoft. Gates settled the matter with DEC out
>>>> of court and immediately hired Cutler as the Chief NT Architect.
>>>
>>> And your evidence for these claims is...
>>
>> I'm not the only one who knows these things. Besides, there are
>> records of DEC having sued Microsoft over this issue. Just because
>> the case was dropped again due to Bill Gates's motion to settle
>> matters out of court - which he did by both paying a large sum of
>> money to DEC and by promising to support DEC's Alpha processor in
>> Windows NT - doesn't mean that the records of the complaints were
>> stricken.
>
> I asked for evidence, not elaborated rumour. e.g. who 'knows' them
> and how, when was the court case, etc.

I am not a lawyer, and even if I were, I am also not a US American.
Both Microsoft and DEC were US American companies, and so the lawsuit
was filed in the USA.

In addition to that, Windows NT has been around since the early 1990s.
That's about twenty years ago. Do you honestly expect me to go scouting
the internet for any records of that?

> That there was design influence is uncontroversial, actual copying of
> code is another matter entirely (and seems pretty implausible on
> purely pragmatic grounds, actually).

I'm not into Microsoft stuff, but I know that they have some sort of
"read-only open source" thing you can subscribe to, where you get the
source code of multiple versions of Windows (and possibly other
Microsoft software) shipped to you on CD/DVD media. (You may read and
study the source code, but you may not modify it, recycle it or
redistribute it.) I have friends who have subscribed to this offer and
who thus got to see the source code, and I know plenty of hackers (in
the true sense of the word) who've coded on VMS.

In addition to all of the above, I myself have not written a single line
of code anymore - with the exception of some Bash and a few other
scripted/interpreted languages - since the mid 1990s, and I've never
learned how to /speak/ C. My programming was limited to sequential
BASIC, Pascal and dBASE IV in both DOS and OS/2, ReXX in OS/2, assembler
and COBOL on UNIX (Sperry OS/3), and a little bit of SQL.

In other words, I'm not well-placed to scrutinize the source code of
either NT or VMS. I'd probably be able to find similarities, but that
doesn't mean that I would understand what the code actually does. Most
of what I know about kernels - and I really am interested in the subject
- comes from reading books about operating system design, both on my own
time and as part of my IT training in the early 1990s. However, as you
could possibly glean from what I wrote higher up, all code written by
myself was exclusively designed for userspace.+

[Martin]

Richard Kettlewell wrote:

> The notion that Windows "insists" on a final ^Z is certainly nonsense,
> including for plain text files.

I agree

>> CP/M itself was loosely inspired by both UNIX and VMS. And NT is a
>> virtually identical copy of the VMS kernel, and still carries a lot of
>> the DOS legacy in it, both from DOS directly and from OS/2, which was
>> devised - by IBM, not by Microsoft - to be the successor of DOS for
>> 80286 and later processors. Here too, Gates pulled the same stunt, i.e.
>> he approached Dave Cutler, one of the authors of the VMS kernel, to
>> write NT. Cutler copied "his" VMS code almost verbatim in NT, and was
>> caught in the act by DEC, who fired him on the spot and sued Microsoft.
>> Gates settled the matter with DEC out of court and immediately hired
>> Cutler as the Chief NT Architect.
>
> And your evidence for these claims is...
>

I think Aragorn went a bit over the top with his wording. Obviously the
target hardware and the software API was different enough that Win NT cannot
be a "virtually identical copy of the VMS kernel".

However, the insider report by Jennifer Edstrom and Marlin Eller
("Barbarians led by Bill Gates", ISBN 0-8050-5754-4) makes the following
statements in chapter 7:

"[...] Cutler was one of the original designers of VMS [...] He would father
the next generation of operating system technology, a system that could
scale from [...] to [...] The operating system Cutler had written at DEC
became the architectural basis for NT. [...]"

There is a lot more about Cutler and the team coming over from DEC, but I
cannot quote it all - sorry, no 10 finger typist here. :p

also chapter 6 is a very interesting read, about a u-boat project moving
Windows from a real mode extender to a protected mode OS (Windows 3.0) which
essentially meant the breakage with IBM and the agreed OS/2 strategy.
Obviously, after that, Win NT was the open declaration of war with IBM.

Martin

[Richard Kettlewell]

Aragorn <str...@telenet.be.invalid> writes:
> Richard Kettlewell conveyed the following to alt.os.linux...
>> Aragorn <str...@telenet.be.invalid> writes:
>>> Richard Kettlewell conveyed the following to alt.os.linux...

>>>> The notion that Windows "insists" on a final ^Z is certainly
>>>> nonsense, including for plain text files.
>>>
>>> Really? Like I said, just because it doesn't show _in_ the editor
>>> doesn't mean that it isn't there.
>>
>> Hexdump of a file created in Notepad:
>>
>> 00000000 62 6c 61 68 20 62 6c 61 68 20 62 6c 61 68 0d 0a |blah
>> blah blah..|
>> 00000010 62 6c 61 68 20 62 6c 61 68 0d 0a |blah
>> blah..|
>>
>> Note the absence of a trailing 1a.
>>
>> Where did you get the idea that the trailing ^Z was mandatory from?
>
> I didn't exactly say that it was mandatory. I said that it was needed,
> for whatever value of "needed".

Hair-splitting. It's not mandatory in any way, it doesn't insist on it
in any way, it doesn't need it.

> And I was rather speaking of DOS than of Windows.

That is simply false. You specifically referred to Windows.

> Could be that it was just a requirement of the editors in DOS -
> EDLIN.COM, EDIT.COM and friends.

My recollection is that the function of ^Z in MSDOS was essentially that
of c_cc[VEOF] in Unix but I don't remember for sure. I don't recall any
trouble with text files copied between MSDOS and other systems apart
from line endings, so even if it did turn up at the end of some files
I'm not convinced MSDOS can really be said to have insisted on it
either.

>> I asked for evidence, not elaborated rumour. e.g. who 'knows' them
>> and how, when was the court case, etc.
>
> I am not a lawyer, and even if I were, I am also not a US American.
> Both Microsoft and DEC were US American companies, and so the lawsuit
> was filed in the USA.
>
> In addition to that, Windows NT has been around since the early 1990s.
> That's about twenty years ago. Do you honestly expect me to go scouting
> the internet for any records of that?

You made the claim, you produce some justification that someone else can
actually assess. It's pretty obvious that you cannot.

--
http://www.greenend.org.uk/rjk/

[Jasen Betts]

On 2012-01-08, Aragorn <str...@telenet.be.invalid> wrote:
> On Sunday 08 January 2012 19:23, Richard Kettlewell conveyed the
> following to alt.os.linux...
>
>> Aragorn <str...@telenet.be.invalid> writes:
>>> On Sunday 08 January 2012 11:03, Jasen Betts conveyed the following
>>> to alt.os.linux...
>>>> On 2012-01-08, Aragorn <str...@telenet.be.invalid> wrote:
>>
>>>>> But... Windoze also insists on an end-of-file character (^Z) _in_
>>>>> the file...
>>>>>
>>>>> You can always trust Microsoft to take something logical and then
>>>>> mess it up so badly that even its mother wouldn't recognize it
>>>>> anymore.
>>>>
>>>> Usually the the end-of-file marker is optional, only a in a few rare
>>>> cases was it needed.
>>>
>>> It was inconsistent, yes. And it was certainly needed for plain text
>>> files, but it didn't always show up in the editor, even if it was
>>> there. Yet another inconsistency.
>>
>> The notion that Windows "insists" on a final ^Z is certainly nonsense,
>> including for plain text files.
>
> Really? Like I said, just because it doesn't show _in_ the editor
> doesn't mean that it isn't there.

Try it some time. use a binary editor (or other means) to create some text files
with no ^Z, open or otherwise read them with different applications,
see what happens.

--
⚂⚃ 100% natural

[Eef Hartman]

Aragorn <str...@telenet.be.invalid> wrote:
>> Windoze insists on having ^M (CTRL-M) characters at the end of each
>> line and often no new line character at the end of the file.
>
> But... Windoze also insists on an end-of-file character (^Z) _in_ the
> file...

The ctrl-Z is a leftover from CP/M (which didn't store file length up
to the byte, so that you needed something to signify end-of-file in
the last 128 bytes) and mostly isn't needed anymore for all programs
since about MS/PC-Dos 3.x
The ctrl-M (carriage return) is a left-over too from CP/Mi, although
Windows still uses it, and from RT/11 before it (where CP/M got a
lot of its inspiration from).
It makes it easy to print/show files as all the carriage control
for _printing_ terminals is already in the file, so needing no
driver actions (raw/cooked in unix/linux).
--
******************************************************************
** Eef Hartman, Delft University of Technology, dept. SSC/ICT **
** e-mail: E.J.M....@tudelft.nl - phone: +31-15-27 82525 **
******************************************************************

[Aragorn]

On Monday 09 January 2012 00:18, Richard Kettlewell conveyed the

following to alt.os.linux...

> Aragorn <str...@telenet.be.invalid> writes:
>> Richard Kettlewell conveyed the following to alt.os.linux...
>>> Aragorn <str...@telenet.be.invalid> writes:
>>>> Richard Kettlewell conveyed the following to alt.os.linux...
>
>>>>> The notion that Windows "insists" on a final ^Z is certainly
>>>>> nonsense, including for plain text files.
>>>>
>>>> Really? Like I said, just because it doesn't show _in_ the editor
>>>> doesn't mean that it isn't there.
>>>
>>> Hexdump of a file created in Notepad:
>>>
>>> 00000000 62 6c 61 68 20 62 6c 61 68 20 62 6c 61 68 0d 0a |blah
>>> blah blah..|
>>> 00000010 62 6c 61 68 20 62 6c 61 68 0d 0a |blah
>>> blah..|
>>>
>>> Note the absence of a trailing 1a.
>>>
>>> Where did you get the idea that the trailing ^Z was mandatory from?
>>
>> I didn't exactly say that it was mandatory. I said that it was
>> needed, for whatever value of "needed".
>
> Hair-splitting. It's not mandatory in any way, it doesn't insist on
> it in any way, it doesn't need it.
>
>> And I was rather speaking of DOS than of Windows.
>
> That is simply false. You specifically referred to Windows.

That may be how I have worded it, and if so, then I apologize for the
innuendo. It was however what I was thinking about.

>>> I asked for evidence, not elaborated rumour. e.g. who 'knows' them
>>> and how, when was the court case, etc.
>>
>> I am not a lawyer, and even if I were, I am also not a US American.
>> Both Microsoft and DEC were US American companies, and so the lawsuit
>> was filed in the USA.
>>
>> In addition to that, Windows NT has been around since the early
>> 1990s. That's about twenty years ago. Do you honestly expect me to
>> go scouting the internet for any records of that?
>
> You made the claim, you produce some justification that someone else
> can actually assess. It's pretty obvious that you cannot.

Whose life depends on it? Why do you need evidence from me over
something that happened almost twenty years ago?

If you really care about the facts, you'll do your own homework,
regardless of what anyone on Usenet says or claims. This is Usenet, not
a courtroom. Freedom of speech, and freedom of opinion.

If on the other hand you don't really care about the facts but you would
happen to have some stakes at play - like a Microsoft paycheck - in
making Microsoft DOS or Windows appear better than they are, then that's
your business. Absence of evidence is not evidence of absence.

Personally, I don't care much for Microsoft, and even less for
propaganda. I am simply reporting on what other people (better equipped
to deal with this issue than myself) have informed me of.

[Aragorn]

On Monday 09 January 2012 09:16, Jasen Betts conveyed the following to

Sorry, I don't run Microsoft software on my systems. On my own
machines, I only ran DOS with Windows 3.0 for about six months back in
1991, and NT 4.0 Workstation for two years between 1997 and 1999. In
between that, I was running IBM OS/2, and I've been using GNU/Linux
exclusively since january 1st, 2000.