More RH questions
Matt McLeod
Is there some way to regenerate RedHat/base/hdlist (or whatever
it is which is used as the package list on an install CD),
to take account of updated packages and new stuff which has
been added?
I'd like to be able to fold the updates and a couple of contrib
packages into our default installation, as that could save a lot
of time in the long run. I'm sure it can be done, as I've seen
people offering CD images with the updates folded in.
Matt
(who, after RH5.x, is quite scared that RH6 is looking like
the least-worst option right now...)
--
The giraffe you thought you offended last week is
willing to be nuzzled today.
Andrew J Cosgriff
> OK, another question.
>
> Is there some way to regenerate RedHat/base/hdlist (or whatever
> it is which is used as the package list on an install CD),
> to take account of updated packages and new stuff which has
> been added?
>
> I'd like to be able to fold the updates and a couple of contrib
> packages into our default installation, as that could save a lot
> of time in the long run. I'm sure it can be done, as I've seen
> people offering CD images with the updates folded in.
Do a search for kickstart on www.redhat.com - I saw mention of this
(there's a program called "genhdlist" I think) on one of the pages
there (I forget whether it was the kickstart howto or the RedHat
manual itself).
--
Andrew J Cosgriff <a...@bing.wattle.id.au> exterminate all rational thought !
zodiac radio CD father purple
Matt McLeod
>Do a search for kickstart on www.redhat.com - I saw mention of this
>(there's a program called "genhdlist" I think) on one of the pages
>there (I forget whether it was the kickstart howto or the RedHat
>manual itself).
Thanks for that. As I've mentioned elsewhere, I've had to do all
of this in a real hurry, hence all the questions which probably
seem rather dense.
Matt
--
"If you're a real good kid, I'll give you a piggy-back ride on a buzz-saw."
-- W. C. Fields
Matt McLeod
appear to have gotten *most* of this stuff going now.
In case anyone is interested, those Vectras suck.
If you stick the kickstart file on the floppy, they'll
go out and find their network settings (DHCP or BOOTP)
quite happily. But they won't do it before loading the
kickstart file.
Final thing to do is to figure out the post-install stuff.
I've got my quickie postinstall script sitting in with
the kickstart files, but sticking:
mount nfshost:/export/linux/kickstart /mnt
/mnt/post/postinstall
into the %post section will mount the stuff OK, but I get
"Permission denied" when trying to run /mnt/post/postinstall
(which is rx for all, so it *should* be OK). Running
it by hand from VC2 works fine, though.
Anyway, that's a relatively minor issue. Worse case I'll
just roll all that up into an rpm.
Tomorrow the first of the kickstarted machines gets
stress-tested^Wused by a "real user".
Matt
Bryan C. Andregg
>Final thing to do is to figure out the post-install stuff.
>I've got my quickie postinstall script sitting in with
>the kickstart files, but sticking:
>
>mount nfshost:/export/linux/kickstart /mnt
>/mnt/post/postinstall
>
>into the %post section will mount the stuff OK, but I get
>"Permission denied" when trying to run /mnt/post/postinstall
>(which is rx for all, so it *should* be OK). Running
>it by hand from VC2 works fine, though.
NFS and executables can be real bitchy if the UID from the server doesn't map
at all to a UID on the client under Linux.
--
Bryan C. Andregg * <band...@redhat.com> * Red Hat, Inc.
1024/625FA2C5 F5 F3 DC 2E 8E AF 26 B0 2C 31 78 C2 6C FB 02 77
1024/0x46E7A8A2 46EB 61B1 71BD 2960 723C 38B6 21E4 23CC 46E7 A8A2
Nile Evil Bastard
Matt McLeod <m...@labyrinth.net.au> wrote:
:In case anyone is interested, those Vectras suck.
Can I just give that a big "YEAH!!!"
It's suitable for use as a pariah machine, i.e. put NT on it.
I use my old Vectra (very old) to test Win 3.1 antivirus builds on. Hard
part is telling whether a crash is the current build or the system itself.
It's fun deliberately filling it with viruses, though.
Bastard machines. Can't even usefully strip them down.
--
http://netizen.com.au/ http://www.caube.org.au/
ALERT! Australian government attempts to censor Australian internet out of
existence! - See http://www.efa.org.au/ for the gory details ...
Matt McLeod
>On 19 Aug 1999 09:34:36 GMT, Matt McLeod <m...@labyrinth.net.au> wrote:
>>I've got my quickie postinstall script sitting in with
>>the kickstart files, but sticking:
>>
>>mount nfshost:/export/linux/kickstart /mnt
>>/mnt/post/postinstall
>>
>>into the %post section will mount the stuff OK, but I get
>>"Permission denied" when trying to run /mnt/post/postinstall
>>(which is rx for all, so it *should* be OK). Running
>>it by hand from VC2 works fine, though.
>
>NFS and executables can be real bitchy if the UID from the server doesn't map
>at all to a UID on the client under Linux.
The files are owned by root - I anticipated trouble. :-)
I'm loathe to put it all into the kickstart file. This portion
shouldn't change from machine to machine, while other parts almost
certainly will - plus it'd be nice to keep it out of the kickstart
file so I can run it seperately when needed.
Still, once I got it mostly working, it seems pretty painless.
Way easier than installing from CD.
Matt
--
"Bastard's the name, but you can call me Right Bleedin'"
Russ Allbery
> Bryan C. Andregg write:
>> NFS and executables can be real bitchy if the UID from the server
>> doesn't map at all to a UID on the client under Linux.
> The files are owned by root - I anticipated trouble. :-)
Which is, of course, the other UID that NFS can get really bitchy about...
--
Russ Allbery (r...@stanford.edu) <URL:http://www.eyrie.org/~eagle/>
Russ Allbery
> Russ Allbery <r...@stanford.edu> wrote:
>> Which is, of course, the other UID that NFS can get really bitchy
>> about...
> Hm. I'd thought that it was a case of root usually being squashed? The
> script (and all it's little friends) are rx for all, so I'd have
> expected that even root on another host could run them.
> But since this is the first time I've tried to do NFS voodoo (we use NFS
> a *lot*, but it was pretty much all in place before I got here), I'm
> assuming that the chances of my being wrong about that are pretty good.
I'm totally not an NFS person. I just know that NFS servers and clients
do weird things with remapping root UIDs and that if remapping UIDs is a
problem, switching to root ownership may not fix it. But if it's rx for
all, I'd agree with you and think anyone should be able to run it.
Matt McLeod
>On 19 Aug 1999 09:34:36 GMT,
>Matt McLeod <m...@labyrinth.net.au> wrote:
>
>
>Can I just give that a big "YEAH!!!"
>
>It's suitable for use as a pariah machine, i.e. put NT on it.
If I put NT on it, I'll be expected to support NT. So
there's no way I'm putting NT on it.
>Bastard machines. Can't even usefully strip them down.
Well... You could nick the disks and the NIC, but probably
not a lot else.
This is possibly why VBC's like them?
Oddly, the Kayaks are perfectly happy to Kickstart
properly - even easier than the mongrel I'd been testing
with, as the initial BOOTP request works fine. Started
doing one yesterday, went to answer a co-worker's question,
came back expecting to have to hit enter once or twice,
but it was already halfway installed.
Matt
Matt McLeod
>Matt McLeod <m...@asac.ericsson.se> writes:
>> Bryan C. Andregg write:
>
>>> NFS and executables can be real bitchy if the UID from the server
>>> doesn't map at all to a UID on the client under Linux.
>
>> The files are owned by root - I anticipated trouble. :-)
>
Hm. I'd thought that it was a case of root usually being
squashed? The script (and all it's little friends) are
rx for all, so I'd have expected that even root on another
host could run them.
But since this is the first time I've tried to do NFS voodoo
(we use NFS a *lot*, but it was pretty much all in place before
I got here), I'm assuming that the chances of my being
wrong about that are pretty good.
Matt
Nile Evil Bastard
Matt McLeod <mjm+u...@labyrinth.net.au> wrote:
:On 20 Aug 1999 05:24:10 GMT, Nile Evil Bastard <nil...@iname.com> wrote:
[HP Vectra]
:>Bastard machines. Can't even usefully strip them down.
:Well... You could nick the disks and the NIC, but probably
:not a lot else.
The disks, yes. Does the NIC actually work with any other hardware on
earth? Have you tried this?
The long skinny novelty power-supply has amusement value, I suppose ...
:This is possibly why VBC's like them?
eeuwww.
My internal-network machines are generic shitboxes bought when this was a
tiny company. My corporate-network machine (the PIII-450 I use for email
and very little else[1]) is a Dell, moderately-proprietised I think.
[1] I put Paint Shop Pro 6 on it, which sucks more than 5 sucked more than
4. Anyone got a 32-bit build of 3 lying around? [2]
[2] oh, newsgroups are supposed to have topics, aren't they.
Matt McLeod
>On 22 Aug 1999 06:39:36 GMT,
>Matt McLeod <mjm+u...@labyrinth.net.au> wrote:
>:On 20 Aug 1999 05:24:10 GMT, Nile Evil Bastard <nil...@iname.com> wrote:
>
>[HP Vectra]
>
>:>Bastard machines. Can't even usefully strip them down.
>
>:Well... You could nick the disks and the NIC, but probably
>:not a lot else.
>
>The disks, yes. Does the NIC actually work with any other hardware on
>earth? Have you tried this?
Nope. but my understanding is that (at least in the ones we've
got) they're simply 3Com NICs - nothing particularly special.
>The long skinny novelty power-supply has amusement value, I suppose ...
>
>:This is possibly why VBC's like them?
>
>
>eeuwww.
>
>My internal-network machines are generic shitboxes bought when this was a
>tiny company. My corporate-network machine (the PIII-450 I use for email
>and very little else[1]) is a Dell, moderately-proprietised I think.
I rather wish we could just order from wherever we like -- at least
then we could give a fairly precise spec and expect to actually
get that.
One of the arguments being used in favour of PCs over
workstations is that the hardware is generic. I can only
assume that the manager(s) in question haven't looked too
closely at the hardware we're getting...
Matt