Preventing SuperUser Access over telnet session SCO UNIX (v5.0.5)
Elsa Montgomery
Basically I wish to restrict root access to the console only......
Thanks
Elsa
Tony Lawrence
> Does anyone know how to prevent root access over a telnet session.
> Basically I wish to restrict root access to the console only....
from
http://pcunix.com/SCOFAQ/scotec1.html#rootlogin
How can I prevent root logins over telnet?
First, you can limit root logins to a specific device by adding an entry
to /etc/default/login:
CONSOLE=/dev/tty01
restricts root to the ALT-F1 multiscreen only. Unfortunately, there
doesn't seem to be any way to add more than one device with CONSOLE, so
that locks root out of tty02 and all the rest.
Less secure, but still useful, is to test within /etc/profile to boot
root out of any but the multiscreens. Code to do that might look like:
IAM=`who am i | cut -d" " -f1`
TTY=`tty`
if [ $IAM = "root" ]
then
case $TTY in
/dev/tty[0-1][0-9]) : ;;
*) exit 0;;
esac
fi
But that simplistic form can cause problems with single user mode login.
The 5.0.6 release adds a REMOTE_ROOT_OK keyword that can be used in
/etc/default/login that *will* allow network access while when CONSOLE
is set! See "man login" (5.0.6).
--
Tony Lawrence
SCO/Linux Support Tips, How-To's, Tests and more: http://pcunix.com
Matt Schalit
>CONSOLE=/dev/tty01
>
>
>restricts root to the ALT-F1 multiscreen only. Unfortunately, there
>doesn't seem to be any way to add more than one device with CONSOLE, so
>that locks root out of tty02 and all the rest.
This is another in a long list of reasons to move to Open Unix 8,
which allows:
CONSOLE=/dev/console,/dev/vt00-/dev/vt14
so that root can come in where root needs to. But OSR5 people are
determined not to migrate, forcing SCO and Caldera to support, what,
five OS's these days? My oh my.
Regards,
Matthew
Tony Lawrence
I don't know if it's "determined not to migrate"; I think it's more FUD
(fear uncertainty and doubt)- will my app run, what changes will I need
to make, maybe I should just give up and go to NT etc..
Stephen M. Dunn
$Does anyone know how to prevent root access over a telnet session.
$Basically I wish to restrict root access to the console only......
Those are two slightly different needs if there are any serial
terminals or modems present ...
Anyway, look at the man page for login(C) and see the section on
how the CONSOLE line in /etc/default/login works.
--
Stephen M. Dunn <ste...@stevedunn.ca>
>>>----------------> http://www.stevedunn.ca/ <----------------<<<
------------------------------------------------------------------
Say hi to my cat -- http://www.stevedunn.ca/photos/toby/
Matt Schalit
>Matt Schalit wrote:
>
>> This is another in a long list of reasons to move to Open Unix 8,
>> which allows:
>>
>> CONSOLE=/dev/console,/dev/vt00-/dev/vt14
>>
>> so that root can come in where root needs to. But OSR5 people are
>> determined not to migrate, forcing SCO and Caldera to support, what,
>> five OS's these days? My oh my.
>
>
>I don't know if it's "determined not to migrate"; I think it's more FUD
>(fear uncertainty and doubt)- will my app run, what changes will I need
>to make, maybe I should just give up and go to NT etc..
Yes, you're right, Tony. FUD is more likely the reason. But after
nearly five years of UnixWare 7 and Open Unix 8, the fact that people
like you and the other regs are still based on OSR5 from what I can
tell sort of amazes me. That fact that so many of you are heard to
say, "It'd be a good idea if you upgrade your old 5.0.[245] to 5.0.6"
and not "to Open Unix 8 if at all possible" seems to reinforce the idea.
I'm just afraid that a fractured Unix user base will leave Caldera
with only the option of putting Unix out to pasture.
As Bill V. mentioned, FreeBSD runs great, and it has the Linux Binary
compatability package that doesn't take up 5 GB like the LKP.
Though I guess as long as there's Rite Aid (runs OSR5 via satellite links)
and fast food, OSR5 will be in demand.
Regards,
Matthew
Bill Vermillion
Matt Schalit <msch...@pacbell.net> wrote:
>On Fri, 08 Mar 2002 20:55:54 GMT, Tony Lawrence <to...@pcunix.com> wrote:
>
>>Matt Schalit wrote:
>>
>
>>> This is another in a long list of reasons to move to Open Unix 8,
>>> which allows:
>>> CONSOLE=/dev/console,/dev/vt00-/dev/vt14
>>> so that root can come in where root needs to. But OSR5 people are
>>> determined not to migrate, forcing SCO and Caldera to support, what,
>>> five OS's these days? My oh my.
>>I don't know if it's "determined not to migrate"; I think it's
>>more FUD (fear uncertainty and doubt)- will my app run, what
>>changes will I need to make, maybe I should just give up and go to
>>NT etc..
>Yes, you're right, Tony. FUD is more likely the reason. But after
>nearly five years of UnixWare 7 and Open Unix 8, the fact that
>people like you and the other regs are still based on OSR5 from
>what I can tell sort of amazes me. That fact that so many of you
>are heard to say, "It'd be a good idea if you upgrade your old
>5.0.[245] to 5.0.6" and not "to Open Unix 8 if at all possible"
>seems to reinforce the idea. I'm just afraid that a fractured Unix
>user base will leave Caldera with only the option of putting Unix
>out to pasture.
I just moved my last Xenix client and fixed two others going into
the Y2K event. Customers just don't want to upgrade when things
are working just as they need them. We've seen people here saying
they still have hundreds of sites on Xenix.
The upgrade from 5.0.0 to more modern makes a lot of sense because
it fixes all the bugs in the original. 5.0.0 was pretty bad in
many areas. Many things didn't get really fixed until the 5.0.4
and I remember one supplement [was it tcp based] that took about 6
tries to get right.
An upgrade is pretty painless. With proper caveats an in situ
upgrade does work - though most here warn against it. Using JPR's
scripts make it pretty painless.
However moving from one OS to another is not quite as painless.
Moving form Xenix to Unix 2.3.4.2 - the OSR5 to UW2/7, OU8 is more
work than an upgrade in an OS tree.
Getting used to SysV4 ways of doing things from the SysV3 was take
a bit of getting used to.
But it seems the customers are the biggest holdouts. A year ago I
was in fixing something for a 2.3.4.2 customer. They did not want
to upgrade because the HW was working so well - NCR Micro-Channel
and an OS upgrade meant getting rid of their super-reliable NCR.
No matter how much sense it makes to upgrade sometimes customers
won't have it if it can be avoided at all.
>As Bill V. mentioned, FreeBSD runs great, and it has the Linux Binary
>compatability package that doesn't take up 5 GB like the LKP.
The largest run time loadable modules are 400K and there are only
six that are over 100K. nfs, smbfs, the Alteon gigabit NIC
driver, the Qlogic scsi/fibrechanel driver, the usb module,
and vinum [their logical disc volume manager]. But the generic
kernel is now up to 3.5MB and mine is a bit over 4MB [I've added a
few things] but overall it tries to avoid bloat.
Most of what I'm using FreeBSD for is in a pure network/ISP
environment. There definately is more support for Linux out there.
It all boils down to the base thought that seems to get overlooked
so often. Find the application that best does the job you need and
then find the OS which supports that best.
I see too many who want to move to Linux because 1) it's cheaper
than SCO/Caldera or 2) it's not Micrsoft. First time base line
costs are often quite deceptive. Popular press can mislead far too
many who make the decisions.
FreeBSD [or any of the BSD variants] wont be the OS of choice for
many applications. But it works well for me. When the Lion
worm/virus gave root access to many Linux systems all it did was
stop DNS in the BSD. That's important in an ISP environment, but
it's all specialized.
I don't think many people realize just how deep the knowledge at
Caldera goes. Ransom Love - CEO of Caldera - used to be at
Sanyo/Icon when they made what to >me< was one of the most
impressive Unix machine on the market.
It could run BSD, SysV, VMS and MS-DOS >>CONCURRENTLY<< - and that
is no mean feet. Thg biggest system had five Motorola 68020s
[maybe 60840s] and an 80286 - all doing dedicated task. It had no
real memory as we think of it and each processor had up to 8MB of
cache for the proccessor. That was a lot for it's time.
It seems to me Caldera has the potential to become the best.
It's just that the popular press always seems to equate
Linux with RedHat.
But to run/recommned and os just because some else is using it is
not the best reason unless your needs are very similar.
>Though I guess as long as there's Rite Aid (runs OSR5 via satellite
>links) and fast food, OSR5 will be in demand.
How is the performance over satellite. There are inherent delays
in that technology. If you want to see an interesting ascii
diagram lookup rfc 3077 on tunneling for UDL - Unidirectional Data
Links - showing how you tunnel under the network layer to emulate
bidirectional links. I'm constantly amazed by people who get these
ideas and then can implement them.
Bill
--
Bill Vermillion - bv @ wjv . com
Matt Schalit
>It all boils down to the base thought that seems to get overlooked
>so often. Find the application that best does the job you need and
>then find the OS which supports that best.
Yes I keep forgetting this.
On the topic of getting software for an OS and what
that means for the viability of the OS, I should take
some time and make a reasoned post later on.
>I don't think many people realize just how deep the knowledge at
>Caldera goes. Ransom Love - CEO of Caldera - used to be at
>Sanyo/Icon when they made what to >me< was one of the most
>impressive Unix machine on the market.
>
>It could run BSD, SysV, VMS and MS-DOS >>CONCURRENTLY<< - and that
>is no mean feet. Thg biggest system had five Motorola 68020s
>[maybe 60840s] and an 80286 - all doing dedicated task. It had no
>real memory as we think of it and each processor had up to 8MB of
>cache for the proccessor. That was a lot for it's time.
Please don't stop posting these little tidbits. I love
that stuff. Sometimes I wonder if I troll this malarky
out there if only to get you guys to drop one of these
gems on us :)
>It seems to me Caldera has the potential to become the best.
The best Linux?
They have a dissilusioned their low level customers - the
desktop users and broke college guys who didn't like their
new licensing costs and who put a lot of effort into making
apps, mods, howtos, and web sites for the OS they like.
Without a loyal cheering section, how can they engender
themselves to the masses? Volution? :-)
>It's just that the popular press always seems to equate
>Linux with RedHat.
>
>But to run/recommned and os just because some else is using it is
>not the best reason unless your needs are very similar.
Ok. Fair enough, but Ou8 doesn't have a lot of proponents.
I guess the revolution is going to be up to Caldera to
instigate from the top down with unique, effective products
and brilliant advertising.
Is it cold in here, or is it just me? :-)
>>Though I guess as long as there's Rite Aid (runs OSR5 via satellite
>>links) and fast food, OSR5 will be in demand.
>
>How is the performance over satellite? There are inherent delays
>in that technology.
I'm not sure that I can quantify it well enough as I don't have
sysop access, though my Dad's a pharmacist there. (my Dad? yes
he should have retired long ago. :) )
What he tells me is that they have a handfull of ethernet based terminals
that connect to a barebones, static OSR5 system in the same store that's
the router/gateway to the satellite link. All records, logins, and
database information is kept remotely, back East.
So everything that they do has to go over the satellite. If the back
East machines are not loaded down doing Oracle lookups, the sattelite
link is nearly instantaneous, I'm told. No more than 4 sec. to send
all the data for a Blue Cross confirmation out and get a response.
So the link isn't the bottleneck, but rather it's the database processing
that slows things down when there's a big load.
>If you want to see an interesting ascii
>diagram lookup rfc 3077 on tunneling for UDL - Unidirectional Data
>Links - showing how you tunnel under the network layer to emulate
>bidirectional links. I'm constantly amazed by people who get these
>ideas and then can implement them.
I'll bookmark that one. Thanks. -- Matt
Tony Lawrence
> On Fri, 08 Mar 2002 20:55:54 GMT, Tony Lawrence <to...@pcunix.com> wrote:
>
>
>>Matt Schalit wrote:
>>
>>
>
>
>>>This is another in a long list of reasons to move to Open Unix 8,
>>>which allows:
>>>
>>> CONSOLE=/dev/console,/dev/vt00-/dev/vt14
>>>
>>>so that root can come in where root needs to. But OSR5 people are
>>>determined not to migrate, forcing SCO and Caldera to support, what,
>>>five OS's these days? My oh my.
>>>
>>
>>I don't know if it's "determined not to migrate"; I think it's more FUD
>>(fear uncertainty and doubt)- will my app run, what changes will I need
>>to make, maybe I should just give up and go to NT etc..
>>
>
>
> Yes, you're right, Tony. FUD is more likely the reason. But after
> nearly five years of UnixWare 7 and Open Unix 8, the fact that people
> like you and the other regs are still based on OSR5 from what I can
> tell sort of amazes me. That fact that so many of you are heard to
> say, "It'd be a good idea if you upgrade your old 5.0.[245] to 5.0.6"
> and not "to Open Unix 8 if at all possible" seems to reinforce the idea.
> I'm just afraid that a fractured Unix user base will leave Caldera
> with only the option of putting Unix out to pasture.
That's the problem. Let's say Caldera decides tomorrpw to scrap OSR5.
How many current OSR5 customers will switch to OU8? My bet is not very
many. and I think it's pretty obvious Caldera has the same opinion or
they would have scrapped it already.
People like me are driven by our customers. I've already switched to
Linux for my personal use, and I've been selling Linux to new customers
for some time now. But I have had very few customers even express any
interest in moving off OSR5 and of those few who have, not one has even
entertained the notion of OU8- it's NT, NT, NT and one AIX.
Now- I'm hoping that the next release of OU8 will be positioned and be
capable of being a truly painless upgrade from 5.0.x. That's going to
be a tough act to pull off, but if they can do it, maybe there is a hope
of migrating at least some good percentage of OSR5 users over. Without
a painless, near invisible transfer, it's just never going to happen-
those OSR5 systems will slowly be replaced rather than upgraded, and
most of the replacements will not be Unix or Linux.
SCO cooked their own goose when they bought Unixware. That's what put
them into this awful situation, but with the possibility of an OSR5
kernel personality for OU8 we have at least the chance that the goose
will come out of the oven smelling good and be ready to serve (groan!).
But (to push the metaphor right over the limit), these OSR5 users are
like people used to having turkey for Thanksgiving dinner. That goose
better look and taste a lot like turkey or they'll be buying Microsoft
Baked Ham for sure :-)
Bill Vermillion
Tony Lawrence <to...@pcunix.com> wrote:
>SCO cooked their own goose when they bought Unixware. That's what put
>them into this awful situation, but with the possibility of an OSR5
>kernel personality for OU8 we have at least the chance that the goose
>will come out of the oven smelling good and be ready to serve (groan!).
I suspect that SCO might not exist today if it hadn't purchased
Unixware. Larger users needed the power and resouces that a SysV4
gave them. The Unixware SMP was about the best out there in terms
of efficieny. And when data grew where 2GB file sizes became
more common the fact that the filesystem size was expanded on OSR5
didn't help that problem.
There are many large systems from what I can infer from the posts
on this group, who are on the UW/SysV4 based product. Even more on
the Unixware group. If SCO hadn't purchased Unix, then those customers
could have gone to Sun/HP/Compaq/IBM. That would have left SCO with
only smaller users - many of whom will not spend money until they
absolutely must. At that point the base-line OS cost makes many want
to move to the Linux environment. Many look at the base OS cost and
don't see the added expense and re-training it takes for some.
So I'll take the opposite POV that if SCO hadn't bought Unix
they might not exist today, or existed long enough for Caldera to
have purchased them. So many companies that sold Unix in the past
either went out of business or dropped Unix as a SW product
and continued their other business.
That included branded Unix from enties such as MicroPort,
Interactive, System, Kodak [Did their Unix come in a yellow box.
Their audio tape came in beige boxes], iNTEL, Esix, et al. All in
the iNTEL based world
There was at one time over 70 variants of Unix all with their own
name. Now you can count on the fingers of one hand the survivors.
Many of those were not iNTEL PC-BUS based but there were several
who used iNTEL processors in their own architecture and weren't
hampered by the HW limitations.
And if SCO hadn't purchased Unixware who knows what would have
happened. Novell's CEO, Bob ???, wanted to diversify so off they
went buying WordPerfect, Unix, et al, and found that they could
not be a mini-MS - a company that supplies all pieces for an
envirornment and began spinning of the pieces/parts.
I will say that at least they were accessible. I had problems with
the local UW 1.x/2.0 vendor - as they were more intested in the
Novell network side - so I sent email to him - and did get handed
off to the right people.
Sun paid about $50 million for a one time perpetual license for
part of the code to SCO when the purchase went down - so they
didn't have to pay royalties on everything they sold.
What if Sun had purchased Unixware, or HP, or ???. SCOs biggest
problem - and part of Caledera's problem [as I see it] is that it
is extremely difficult to be a SW only operation. All the other
independant Unix SW only houses eventually folded leaving SCO alone.
Because of such things as the Lotus suit SCO dropped their
spread-sheet. Lyrix couldn't hold it's own against the push from
MS for 'pretty' word-processors. All the things SCO used to do,
cross-compilers, porting, etc., just slowly went away until they
were left as just an OS company. Even MS isn't that - though we
think of them that way. Buying an OS from MS is just the first
step in a large pile of dollars going their way, as anyone knows.
Just look at the application upgrades for Office and how that is a
must for many.
You see posts from people still using OSR5. That came out
in 1995. What else came out in 95. Win95 [late - in October]
But MS had dropped support of Win95 as of the announcement last
week. SCO didn't drop support of Xenix until 1998 - almost 10 years
after they first announced it was going away and 8 years after they
announced support would stop in 1990. [I hope I'm remembering the
time line right].
You can try to be all things to all users/customers - but that can
cost you more than if you said "It's time to upgrade - we no longer
support this OS - if you choose to remain on the old one then you
are on your own".
There are still people using Window for Workgroups 3.11 - but they
are certainly in the minority.
The MS OS plus applcation support - where the applications become
one of the largest parts of your investment reminds me somewhat of
the original Gillette concept of giving away the razor and selling
the razor blades. MS has a lot of 'blades' to sell - but not many
Unix vendors do.
>But (to push the metaphor right over the limit), these OSR5 users are
>like people used to having turkey for Thanksgiving dinner. That goose
>better look and taste a lot like turkey or they'll be buying Microsoft
>Baked Ham for sure :-)
I suspect that the smaller users are happy with the OSR5 - but how
many people can be limited by the 2GB file sizes inherent in a V.3
system. Some people are resistant to change and that's where your
metaphor fits. It does take time, effort and money to change.
But if at least try to make small changes along the line you don't
come to the day when you must change and the costs are so high you
can't afford to do it.
The amount of data that users want to keep is exploding. A few
years ago 1GB disk were though to be big. Now you can buy
NAS devices that are well over 1TB at fairly reasonable prices.
These would not have existed had the need not been there.
And TB databases are starting to look small for bigger vendors -
and that means the smaller ones will follow.
And big systems are growing at rates that were unimaginable just 2
years ago.
Sears is reworking their systems to be able to tie all of a
customers purchases together to be able to target specific things
the customer might be doing. Current they can track the sales to
the item. What they will be doing is taking individual sales
and mapping them for special offers.
If a customer buys certain things they might be going on a vacation.
Their current database is 70TB - but with the new appoach they
exepct to cross the 1PB [one petabyte] boundary by 2004. That is not
very far in the future. I saw a Petabyte tape backup system over 3
years ago. So this fits the old saw of 'data expands to fill all
useable storage'.
I remember when a 1GB file was large - but when you start having
advertising agencies send files to print shops - then 50MB is
normal. I had to open up the default on the mail handler to let
that go through.
Granted that Xenix still works for many people but for most the file
size and memory size limits make it unworkable.
It seems that in the past everytime SCO changed something a great
many complained. The move to the 3.2 Unix was slow - and people
resisted change. The same thing happened with OSR5 and UW2 and UW7
and now OU8. It almost seem to be a prerequisite here, if something
changes, complain about it.
At least we don't see so many jumping on a poster for asking
Unixware questions here as we used to a year or so ago. Even though
this group is comp.unix.sco.misc users saw the SCO on the Unixware
box and assumed 'this must be the place'.
Newsgroup descriptions show that Unixware is a Novell product.
Perhaps it's time for a renaming. comp.os.calder.[productname]
Tony Lawrence
> In article <3C8B2E16...@pcunix.com>,
> Tony Lawrence <to...@pcunix.com> wrote:
>
>
>>SCO cooked their own goose when they bought Unixware. That's what put
>>them into this awful situation, but with the possibility of an OSR5
>>kernel personality for OU8 we have at least the chance that the goose
>>will come out of the oven smelling good and be ready to serve (groan!).
>>
>
> I suspect that SCO might not exist today if it hadn't purchased
> Unixware. Larger users needed the power and resouces that a SysV4
> gave them. The Unixware SMP was about the best out there in terms
> of efficieny. And when data grew where 2GB file sizes became
> more common the fact that the filesystem size was expanded on OSR5
> didn't help that problem.
Everybody always thinks the large users are the important ones, and they
almost never are. Worse, paying attention to their needs is always
easy, eminently sensible and usually the worst thing you can do. The
"follow the mastodons" philosophy has killed many a company- yet people
keep doing it time after time. http://pcunix.com/Books/innovdil.html
If you are a mastodon, you need mastodon sized customers. When you are
a kitty cat, those big hunks of meat are obviously out of your league.
Caldera isn't much more than a good sized tiger and should be hunting
smaller game.
..
> At least we don't see so many jumping on a poster for asking
> Unixware questions here as we used to a year or so ago. Even though
> this group is comp.unix.sco.misc users saw the SCO on the Unixware
> box and assumed 'this must be the place'.
No, but JPR still grouses now and then..
>
> Newsgroup descriptions show that Unixware is a Novell product.
> Perhaps it's time for a renaming. comp.os.calder.[productname]
I wouldn't. I'd bet at least even money that Caldera will not be the
ultimate owner of this stuff. Leave it as unix, because that's what it
is. As the checkered history of Unixware shows, owners come and go..
assuming it survives at all.
Just an ironic note re pursuing the big boys:
If SCO and it's various licensees had paid attention to the desires of
a young Finnish student for a reasonable cost Unix OS, Linux never would
have happened. Greed and blindness is always what causes these
unexpected out of nowhere things to unseat you. But of course personal
use licenses were too unimportant, too small time, and why would you
want to leave money on the table anyway- not until there was no other
choice, not until the writing was writ so large that no one could ignore
was any move made in that direction.
As I've said before, even right now may not be too late. If OSR5 is
dead anyway, why not release source? It's entirely possible that its
stability might attract a substantial base of interested developers.. oh
but why bother- I doubt anyone in power can see the intelligence of that..
Bill Vermillion
>> In article <3C8B2E16...@pcunix.com>,
>> Tony Lawrence <to...@pcunix.com> wrote:
>>>SCO cooked their own goose when they bought Unixware. That's what
>>>put them into this awful situation, but with the possibility of
>>>an OSR5 kernel personality for OU8 we have at least the chance
>>>that the goose will come out of the oven smelling good and be
>>>ready to serve (groan!).
>> I suspect that SCO might not exist today if it hadn't purchased
>> Unixware. Larger users needed the power and resouces that a SysV4
>> gave them. The Unixware SMP was about the best out there in terms
>> of efficieny. And when data grew where 2GB file sizes became
>> more common the fact that the filesystem size was expanded on OSR5
>> didn't help that problem.
>Everybody always thinks the large users are the important ones,
>and they almost never are. Worse, paying attention to their
>needs is always easy, eminently sensible and usually the worst
>thing you can do. The "follow the mastodons" philosophy has
>killed many a company- yet people keep doing it time after time.
>http://pcunix.com/Books/innovdil.html
I'll have to check that out. If you are innovative and good at it,
the Microsoft will swallow you up. The way MS destroyed GO by
announcing OSes that weren't released, etc. was pretty bad. They
even bragged later that it only cost them $4 million to sink a $70
million company. GO had handwriting recognition in a small laptop
style computer. State Farm was going to order 150,000 - if I
remember the figure correctly. Then MS announced "Pen-Computing"
[Wasn't that the OS variant] and spread all sort of FUD until GO
was gone. Documented in the book "Start Up".
Depending on your product line large users can be the important
ones. Several companies who were broad-based - eg small users to
huge user - have taken a close look and said "we want big
customers"
A classic case is GE. They slowly spun off all consumer items.
They first got rid off all their small appliances by selling them
to Black&Decker. You know - the things like the under the counter
coffee-pots, etc. Then they sold the brand name RCA to Thompson in
France and the right to use the GE name on TV to Thompson for a
period of 10 years. A great many never knew that. No more
transitor radios, cassette decks, etc with a GE name.
They decided they wanted big customers. If I thought awhile I could
probably come up with some that did the opposite - and decided
their best approach was small-prices with millions of customer
sales. For GE it made more sense to sell $1,000,000+ diesel
locomotives and even pricier jet plane engines than it did to sell
$300 TV set.
In the computer world we see the larger database companies
targeting larger companies - as at the price for Oracle small
companies cant afford it - and probably don't need it.
There are very few compnaies that cater to the broad range anymore.
Sony comes to mind with things from $1.00 batteries to
multi-hundred thousand dollar electronics things for TV and
broadcast. The Petabyte tape drive I mentioned was Sony. ISTR the
base price started at $95,000. One drive and one cabinet.
>If you are a mastodon, you need mastodon sized customers. When you
>are a kitty cat, those big hunks of meat are obviously out of your
>league. Caldera isn't much more than a good sized tiger and should
>be hunting smaller game.
They seem to have some pretty good sized customers according to the
financial report post to the announce list in December.
==========
During this quarter we continued to provide trusted business solutions
and services to our worldwide customers including McDonalds, CVS
Pharmacy, Rite Aid Pharmacy, Eckerd, Shoppers Drug Mart, NASDAQ,
Goodyear, BMW, Michaels, Tyson Chicken, Publix Super Markets, Polish
Ministry of Finance, The China Post, Ministry of Administration in
Korea, and the Savings Bank of Russia, said Love. We are also seeing
an increase in the number of software platform providers who are
certifying their products on Caldera Open UNIX 8 including such
companies as Informix, Oracle, Progress and Sybase.
==========
There are some decent sized companies in that mix.
>> At least we don't see so many jumping on a poster for asking
>> Unixware questions here as we used to a year or so ago. Even though
>> this group is comp.unix.sco.misc users saw the SCO on the Unixware
>> box and assumed 'this must be the place'.
>No, but JPR still grouses now and then..
And you've stopped - good. {ISTR that you used to do this - if I'm
mistaken I apologize}
>> Newsgroup descriptions show that Unixware is a Novell product.
>> Perhaps it's time for a renaming. comp.os.calder.[productname]
>I wouldn't. I'd bet at least even money that Caldera will not be
>the ultimate owner of this stuff. Leave it as unix, because that's
>what it is. As the checkered history of Unixware shows, owners come
>and go.. assuming it survives at all.
Looking at what I wrote there are two approaches. Currently it's
comp.unix.sco.misc and a change/addition with the various pieces
under Caldera might be good. OTOH the comp.os has such notable
entries as linux, ms-xxxx, solaris, hp-ux, vms, et al.
Since SCO no longer exists by name I would think it would have to
change at some time.
>Just an ironic note re pursuing the big boys:
>If SCO and it's various licensees had paid attention to the desires
>of a young Finnish student for a reasonable cost Unix OS, Linux
>never would have happened.
Or by extension if Prentice-Hall and almost forcing you to buy the
book by Tannenbaum and forcing the target to the educational market
then perhap Linux would have not put the effort into Linux. As I
recall it was Minix that inspired him. Minix first saw light of day
in 1986.
>As I've said before, even right now may not be too late. If OSR5
>is dead anyway, why not release source? It's entirely possible
>that its stability might attract a substantial base of interested
>developers.. oh but why bother- I doubt anyone in power can see the
>intelligence of that..
They could release only part of the source as there are far too
many contingent copyrights involved. Have you ever looked at the
complete copyrights list? Gawd is that long. You surely wouldn't
have a fully functioning OSR5 without all the other pieces.
And of course you could kiss the Xenix compatibility goodbye as I
think it would take more than just a cold day for MS to turn loose
any code they own - and they do have the Xenix and the trademark on
that name.
At least Caldera has put all the previous source out there. Doug
Michaels first let the 'ancient unix' be licensed but it was
$100 and you almost had to have a PDP/Vax to run it. Now much more
is available and the license is free.
Saying that - most of that is good for experimenters and I suspect
all the add-in copyright material in OSR5 would be harded to
extricate than the process that UCB went through stripping the AT&T
code from the Berkeley Software Distribution. The only way that
finally got out was the counter-suit against AT&T by The Regents -
because AT&T had taken BSD code and stripped the copyright
statements. So they agreed to let BSD become 'free' and both sides
dropped their suits.
Who knows what we will see by next year.
Bill
Jean-Pierre Radley
|
| Everybody always thinks the large users are the important ones, and they
| almost never are. Worse, paying attention to their needs is always
| easy, eminently sensible and usually the worst thing you can do. The
| "follow the mastodons" philosophy has killed many a company- yet people
| keep doing it time after time. http://pcunix.com/Books/innovdil.html
|
| If you are a mastodon, you need mastodon sized customers. When you are
| a kitty cat, those big hunks of meat are obviously out of your league.
| Caldera isn't much more than a good sized tiger and should be hunting
| smaller game.
It's suberbly ironic that Clay Christensen, author of "The Innovators
Dilemma" which you so admire, was the keynote speaker at SCO FORUM 2000.
I imagine that SCO probably paid him some honararium, nevertheless SCO
paid little attention to his words; in due course, there came a FORUM
2001 -- but it was hosted by Caldera.
| As I've said before, even right now may not be too late. If OSR5 is
| dead anyway, why not release source? It's entirely possible that its
| stability might attract a substantial base of interested developers.. oh
| but why bother- I doubt anyone in power can see the intelligence of that..
Why do you call it dead, when OSR 5.0.7 is in the works?
--
JP
Bob Meyers
"Elsa Montgomery" <elsa.mo...@travel2.com> wrote in message
news:c405f8b3.0203...@posting.google.com...
> Does anyone know how to prevent root access over a telnet session.
> Basically I wish to restrict root access to the console only......
>
Yes, I know the perfect method! Change root's password so only you know it,
and then only login as root on the console.
Gee, I'll never again apologize for topic drifting. I see us old men are on
the park bench again.
(But I think I will check out that book)
Tony Lawrence
>
>>>At least we don't see so many jumping on a poster for asking
>>>Unixware questions here as we used to a year or so ago. Even though
>>>this group is comp.unix.sco.misc users saw the SCO on the Unixware
>>>box and assumed 'this must be the place'.
>>>
>
>>No, but JPR still grouses now and then..
>>
>
> And you've stopped - good. {ISTR that you used to do this - if I'm
> mistaken I apologize}
Goodness no- I'm the one who was always yelling at JPR and the others
who didn't want UW related posts here!
Tony Lawrence
> Tony Lawrence propounded (on Sun, Mar 10, 2002 at 07:29:49PM +0000):
> |
> | Everybody always thinks the large users are the important ones, and they
> | almost never are. Worse, paying attention to their needs is always
> | easy, eminently sensible and usually the worst thing you can do. The
> | "follow the mastodons" philosophy has killed many a company- yet people
> | keep doing it time after time. http://pcunix.com/Books/innovdil.html
> |
> | If you are a mastodon, you need mastodon sized customers. When you are
> | a kitty cat, those big hunks of meat are obviously out of your league.
> | Caldera isn't much more than a good sized tiger and should be hunting
> | smaller game.
>
> It's suberbly ironic that Clay Christensen, author of "The Innovators
> Dilemma" which you so admire, was the keynote speaker at SCO FORUM 2000.
> I imagine that SCO probably paid him some honararium, nevertheless SCO
> paid little attention to his words; in due course, there came a FORUM
> 2001 -- but it was hosted by Caldera.
That is ironic.
>
> | As I've said before, even right now may not be too late. If OSR5 is
> | dead anyway, why not release source? It's entirely possible that its
> | stability might attract a substantial base of interested developers.. oh
> | but why bother- I doubt anyone in power can see the intelligence of that..
>
> Why do you call it dead, when OSR 5.0.7 is in the works?
It's dead. I don't care if money has already been put in a trust fund
for version 6.0.0, the product is walkng dead and every one of us knows
it. If Unix has any future at all, it's OU8.
That's not to say that there won't be OSR5 installs around for some time
to come. At least, I certainly hope there will be. But the inexorable
march toward larger files, memery, more networking based computing etc.
leave OSR5 in the scrap heap.
Tony Lawrence
Park bench? Oh, sheesh, I must have nodded off again.. I thought we
were at the coffee shop.
Bill Vermillion
>That is ironic.
Perhaps qualifing that slightly to if Unix has a future on iNTEL
based platforms?
Looking at HW sales of Unix servers IBM is #1, Sun #2 and HP #3.
Looking a the SW side it is just reversed Sun, HP and IBM.
Totals in both areas show that these three account for over 75% of
the market.
>That's not to say that there won't be OSR5 installs around for some
>time to come. At least, I certainly hope there will be. But the
>inexorable march toward larger files, memery, more networking based
>computing etc. leave OSR5 in the scrap heap.
Agree completely on those points.
-bill-
>
...
>
> >That's not to say that there won't be OSR5 installs around for some
> >time to come. At least, I certainly hope there will be. But the
> >inexorable march toward larger files, memery, more networking based
> >computing etc. leave OSR5 in the scrap heap.
>
> Agree completely on those points.
>
> Bill
On the other hand: We design and sell systems for medical offices. I doubt we
will _ever_ see a even a 1 GB file and our network connectivity needs are small
- 3-10 internal clients and one highly restricted connection to the internet.
5.0.5/6 just hums in that environment with a pentium II - 400 so why would we
ever want to move to another OS (as long as security can be maintained) ?
--
- bill -
bill at TechServSys {put a period here} com
Brian K. White
"-bill-" <nob...@spamcop.net> wrote in message
news:3c8e0596$0$98142$2c3e...@news.voyager.net...
easy, when they say want t start hooking all that medical information that
comes in the form of various media in to the database so that it is
stored/modified/recalled via the fast efficient computer instead of the
wall of paper files and the room of various odd recordings.
cat scans, mri's, any other 3-d model scans that they dream up, videos of
ultra-sounds, xrays, videos of operations, videos from those scopes they
stick in your various existing and made-to-order orifices via long tubes,
recordings of ekg's, and even the lowly simple 8 1/2 x 11 paper document
scanned in simply for efficiency of day to day work, even if most such
documents in such a field cannot be fully given up like they can in other
fields.
if you are not already offering that, or at least starting to, then you
are slipping behind and you'd better start looking over your shoulder.
Your stuff may be superior, but once the other guy has something you
better have it too or else it will start getting tough to sell.
2 gigs is probably still more than enough for any single file of any type
mentioned so far... it's still a damned long movie for instance even at
vary high image quality, but the system as a whole would need a lot more
beef than without all that stuff, mostly just a boatload of hd space, and
that is certainly doable right now with osr5. as long as you keep the
documents as discrete files and just database the filenames and not the
actual content. sounds obvious but there have been plenty of examples of
databases that hold the actual data for a while now.
actually... even 2 gigs is too small for a single file in some very easily
imaginable cases. A DVD is 4.7 to 8.5 gigs on a side. if you wanted to
film a procedure at full quality it could easily fill or even go beyond
the 8.5 gigs because certainly many procedures end up taking more than a
couple hours.
A mere 2 gigs would be right out the door in that case. And I just
re-thought about cat-scans and 3-d scanning in general... what a collosal
amount of data it must be to scan a person even at low resolution, if you
wanted to keep all pixels instead of some little area of attention! think
of how a simple 8 1/2 x 11 sheet at 300x300 dpi, 16 bit color, is like 65
megabytes, now make the paper as thick as a 300 dpi pixel and make a stack
of sheets as tall as a person, now put two stacks next to each other, or
three, or 4 whatever it takes to get around your waist, .. big file... now
do several such scans a year, now get more than one patient... big unholy
gobs of data, but, if the raw hardware is attainable that can hold and
serve up this kind of data, they will demand it sooner or later, and the
hardware is attainable now.
It doesn't matter that you haven't needed to worry about stuff like this
before, you do now, or you soon will.
I'm feeling it in my (completely unrelated) area. There is no rest for us.
--
Brian K. White -- br...@aljex.com -- http://www.aljex.com/bkw/
+++++[>+++[>+++++>+++++++<<-]<-]>>+.>.+++++.+++++++.-.[>+<---]>++.
filePro BBx Linux SCO Prosper/FACTS AutoCAD #callahans Satriani
Bob Cole
When my father retired from private practice 14 years ago, he had one
computer for EKGs. 14 years before that, he could not imagine why he
would *ever* need or want a computer. The doctor ( and later 2 doctors)
who took over his practice has more than one computer (IIRC). *Ever* is
a very long time.
Brian White gives a lot of examples.
Bob Cole
Bill Vermillion
>> >That's not to say that there won't be OSR5 installs around for some
>> >time to come. At least, I certainly hope there will be. But the
>> >inexorable march toward larger files, memery, more networking based
>> >computing etc. leave OSR5 in the scrap heap.
>> Agree completely on those points.
>On the other hand: We design and sell systems for medical offices.
>I doubt we will _ever_ see a even a 1 GB file and our network
>connectivity needs are small - 3-10 internal clients and one highly
>restricted connection to the internet. 5.0.5/6 just hums in that
>environment with a pentium II - 400 so why would we ever want to
>move to another OS (as long as security can be maintained) ?
And how long will it be before the larger drives and faster
processors make all doctors want to at least take a quick view of
their patients x-rays - instead of having to pull them out of the
files and put them on a box. Pulling them in electronically
would also permit them to enlarge things.
Don't bet on the fact that you won't need large drives :-)
I remember when the ST502 came out. $2500 list price. 5MB.
WOW - what would we do with all that storage. :-) I also remember
members of the group snapping them up at $995 when some refurbs
became available.
For those used to the modern drives you can't appreicated how
delicate these were.
The drive would be in a box probalby 8 x 12 x 6. That box would be
suspended by springs from the corners inside a box that was
at least 2x2x2. That was >standard< shipping method. 1/2 to 1GB
non-operating mode was about the limit.
Now we have laptop drives that can sustain 100G's when operating.
You don't want to be holding that at the time because a human
body fails at about 20G.
Bill Vermillion
Brian K. White <br...@aljex.com> wrote:
>"-bill-" <nob...@spamcop.net> wrote in message
>news:3c8e0596$0$98142$2c3e...@news.voyager.net...
>> Bill Vermillion wrote:
>> >
>> ...
>> > >That's not to say that there won't be OSR5 installs around for
>> > >some time to come. At least, I certainly hope there will be.
>> > >But the inexorable march toward larger files, memery, more
>> > >networking based computing etc. leave OSR5 in the scrap heap.
>> > Agree completely on those points.
>> On the other hand: We design and sell systems for medical
>> offices. I doubt we will _ever_ see a even a 1 GB file and our
>> network connectivity needs are small - 3-10 internal clients and
>> one highly restricted connection to the internet. 5.0.5/6 just
>> hums in that environment with a pentium II - 400 so why would
>> we ever want to move to another OS (as long as security can be
>> maintained) ?
>easy, when they say want t start hooking all that medical
>information that comes in the form of various media in to the
>database so that it is stored/modified/recalled via the fast
>efficient computer instead of the wall of paper files and the room
>of various odd recordings.
>cat scans, mri's, any other 3-d model scans that they dream up,
>videos of ultra-sounds, xrays, videos of operations, videos from
>those scopes they stick in your various existing and made-to-order
>orifices via long tubes, recordings of ekg's,
....
>2 gigs is probably still more than enough for any single file of
>any type mentioned so far... it's still a damned long movie for
>instance even at vary high image quality, ....
Very high image quality take a LOT of space. Baseline broadcast TV
is 14MB/second. MPEG2 encoding for distribution still comes in at
about 35MB/min - and that is standard quality - not the HDTV type
quality - which could be appropriate in medical environments.
However a comment on the mri stuff.
About 10 years ago I saw a technology demonstration from DEC
with voice controller computer displays for real time use in the
operating room.
There was an MRI scan and the scenario was severe head wound.
The surgeon was able to rotate the images in all directions and see
exactly where the bone fragments were inside the brain before he
even picked up his scalpel. It's a pretty amazing world in which
we live.
Simon Hobson
(in message <3C8B2E16...@pcunix.com>):
> That's the problem. Let's say Caldera decides tomorrpw to scrap OSR5.
> How many current OSR5 customers will switch to OU8? My bet is not very
> many.
In our case at work, we would then have the choice of moving to AIX or NT,
and given the cost of AIX (and the hardware to run it on) it would be NT.
Why, because the software we run (Sage Enterprise aka Tetra CS3) runs on
exatly three OSs - OS5, AIX, and NT. NT would be my last choice of platform,
AIX I could not get past the beancounters - personally I'd like to see it on
something else that supports big disk caches but given the noises from Sage I
don't see them putting any effort into porting it to another Unix-based OS,
all their recent stuff is NT only :-(
Simon
Roberto Zini
Well, if one's really interested in running their preferred OS5 applications
and OS5 is no longer available, I think the switch to OU8+OS5 personality is a must.
I mean, with OU8 you can have 3 personalities: the one which supports UW7 programs,
the one (LKP) which supports native Linux binaries and (soon) the OpenServer5
one which (hopefully) should provide a useful, stable and reliable OS5 emulation
(not to mention that you can also have a great Windows emulation under Merge).
From this point of view, I think Caldera's offer will provide one of the most
versatile deployment platform on the market with the choice of several different
"personalities" under which you can run your preferred apps.
Just my 2 cents,
Roberto "I don't work for Caldera - yet :-)" Zini
--
---------------------------------------------------------------------
Roberto Zini email : r.z...@strhold.it
Technical Support Manager -- Strhold Evolution Division R.E. (ITALY)
---------------------------------------------------------------------
"Has anybody around here seen an aircraft carrier?"
(Pete "Maverick" Mitchell - Top Gun)
Simon Hobson
(in message <3C984278...@strhold.it>):
>> In our case at work, we would then have the choice of moving to AIX or NT,
>> and given the cost of AIX (and the hardware to run it on) it would be NT.
>><snip>
> Well, if one's really interested in running their preferred OS5 applications
> and OS5 is no longer available, I think the switch to OU8+OS5 personality is
> a must.
>
> I mean, with OU8 you can have 3 personalities: the one which supports UW7
> programs,
> the one (LKP) which supports native Linux binaries and (soon) the OpenServer5
> one which (hopefully) should provide a useful, stable and reliable OS5
> emulation
> (not to mention that you can also have a great Windows emulation under
> Merge).
Even that is unlikely to be a practical route - it's all down to the (often
mythical) thing called 'support'.
You see, if we run the software on anything but their 'supported' platform
then there is the risk of us finding a problem and the vendor turning round
and saying "tough, unsupported platform, we aren't going to even look at the
problem" - not a risk that we could take. As it is, both the vendor and the
dealer are all too ready to point the finger at something else if they can !
IFF the vendor says that the software is supported on OU8 then fine, but they
havn't officially stated that OS5.0.6 is supported (only 5.0.5 IIRC) !
Having said that, it's an attractive proposition to me (wap the RAM up to 16G
and use it would be nice !) - it's just getting the vendor to accept
responsibility for anything that is the worry.
Simon
Bill Vermillion
Simon Hobson <si...@no-spam.ccomms.demon.co.uk> wrote:
>On Sun, 10 Mar 2002 9:54:34 +0000, Tony Lawrence wrote
>(in message <3C8B2E16...@pcunix.com>):
>> That's the problem. Let's say Caldera decides tomorrpw to scrap
>> OSR5. How many current OSR5 customers will switch to OU8? My bet
>> is not very many.
>In our case at work, we would then have the choice of moving to
>AIX or NT, and given the cost of AIX (and the hardware to run it
>on) it would be NT.
Given some reports and instances of NT the AIX could be a cheaper
overall solution. While working on a Unix system at a client site
I asked the tech there how long it took to rebuild a totally
crashed NT system. He said it averages about 3 days. Most here on
SCO based systems have been using BackupEdge or LoneTar because the
ability to start restores within minutes of drive replacement.
Those packages are also available of AIX. What matters is the cost
after a year or two years - not how much you have to pay today.
The IBM platforms have dropped in price over the last year or so
and they do build rugged HW.
>Why, because the software we run (Sage Enterprise aka Tetra CS3)
>runs on exatly three OSs - OS5, AIX, and NT. NT would be my last
>choice of platform, AIX I could not get past the beancounters -
Do research to see if you can find some real time figures on such
things as system restores after crashes. Lost time can add up and
if the bean-counter knows his beans he might take that into
consideration. For your sake I hope they aren't 100% bottom-line
oriented.
I was amazed at how long the restores could take. But then again
I just saw a site move their mail from a Unix system to an NT that
took almost a month to get it stable. In *n*x it would have been
two days maximum.
Roberto Zini
Well, I know what you mean.
You're not the first who points the finger against this situation where an
application vendor is not willing to support a kind of "emulated" environment;
I think we could think the same about a windows app running under Merge or
whatever Windows emulator you like.
I point I was trying to make is that __IF__ OpenServer is no longer available
(which is not since 5.0.7 is due to be release soon) and you __HAVE__ to run
a given OS5 application, __THEN__ portable openserver could be the only viable
solution.
Apart from that, I'm with you concerning the "support" of applications running
on the above "environment".
Best,
Roberto