Re: No longer supporting Unixware / Open Server
From: Bill Vermillion (bv_at_wjv.comREMOVE)
Date: 01/31/04
- Next message: Bill Vermillion: "Re: No longer supporting Unixware / Open Server"
- Previous message: J. L. Schilling: "Re: No longer supporting Unixware / Open Server"
- In reply to: Joe Dunning: "Re: No longer supporting Unixware / Open Server"
- Next in thread: Brian K. White: "Re: No longer supporting Unixware / Open Server"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Sat, 31 Jan 2004 16:05:01 GMT
In article <VyBSb.146218$5V2.778550@attbi_s53>,
Joe Dunning <joedunning1234_removethis@hotmail.com> wrote:
>On Fri, 30 Jan 2004 19:35:01 GMT, Bill Vermillion <bv@wjv.comREMOVE> wrote:
>>In article <cz2Sb.179512$na.292957@attbi_s04>,
>>Joe Dunning <joedunning1234_removethis@hotmail.com> wrote:
>>>On 28 Jan 2004 19:47:26 -0800, Brian K. White <brian@aljex.com> wrote:
>>>>joe@blahblah.invalid (Joe Dunning) wrote in message
>>>news:<soXRb.46152$U%5.240446@attbi_s03>...
>>>>> On 28 Jan 2004 10:42:47 -0800, Brian K. White
>>>>> <brian@aljex.com> wrote:
>>>>> >And people wonder why I refuse to allow linux into the professional
>>>>> >work environments I'm responsible for? Happily though, I happen to
>>>>> >already know that your attitude is merely one of a very vocal but
>>>>> >happily very minor minority.
>>>>> Brian, you have shown that you are merely an SCO zealot.
>>>>google amazon et al, can afford to have full time 24/7 on-site
>>>>IT staff that can immediately deal with the vagaries of life
>>>>with linux,
>>>Linux does not need 24/7 support. If you think that you are
>>>showing your ignorance. On the other hand, I don't see any
>>>SCO installations in Netcraft's list of machines with longest
>>>uptimes.
>>You say Linux does not need 24/7 support. You also say that you
>>don't see any SCO installation on Netcraft's list of machines with
>>longest uptimes. Well I don't see any Linux machines there
>>either.
>That's true, today. But a week or so ago, there were some. I
>also know from personal experience that Linux machines will run
>for long times at high CPU loads without crashing or the need
>for reboots.
I was surprised to see Sun and a few others missing myself and it
was copletely BSD variants. But since Netcraft is for showing what
web servers are up most of the time, those who run them most are
going to use the ones that work best for that.
>>I support several OSes including a dwindling number of SCO clients.
>>But personally I'd never put SCO up as primarily a web server where
>>it would even be in running for being on one of those lists. It's
>>a good office/business server, but from my POV it is not what I'd
>>consider a server to use in and applications like you'd see on
>>Netcraft. So I don't see the point of bringing in Netcraft and
>>Linux. Am I missing something there.
>Brian claimed the Linux is unstable. I was trying to prove
>otherwise. It looks like my evidence did not support my argument,
>today (as mentioned earlier, Netcraft did show some Linux boxes
>in the top 50 only 2 weeks back)
Instability often seems to depend upon several factor, least of
which seems to be the OS [unless we are talking of older items from
Redmond]. Having run Unix or Unix like systems for about 20 years
I definately prefer the BSD approach as opposed to Linux. And I
detest - well maybe destest is a bit strong - the RPM concept and
all the dependancies, but that was only after being spoiled by the
ports systems and the portupgrade and building everything from
sources each time.
>From my POV of the systems I've worked with the absolute best for
upgrading and making sure all dependancies were met, and that
everyting was in sync, was IRIX from SGI. You never got left
hanging and everything was down to a yes or no when making choices
if there were such things as library conflicts.
The nastiest one I had was years ago on an old SCO 3.2.?.?
intalling 3rd party serial driver that required non-standard flags.
It was all supposed to work, but it would fail. And it was one of
those install packages that when it failed it would remove all the
things it installed. But it did more than that. It also
uninstalled part of the link kit so I'd have to reinstall that
before continuing. But that was fixed by the time the first OSR5s
came out.
But admins who have no clue as to what they are doing seem to be
one of the largest reasons for instabiltiy - at least from systems
I've been aound. The other is cheap hardware and when it has a
problem in a *n*x system and I talk to the HW vendor I get "but it
works in DOS".
I inadvertantly got a person fired who gave me that line, because
they had no clue, and the reason it 'worked in DOS' and not in *n*x
was the way the hardware was treated. The former approached the
device from a sofware perspective, and the latter from the hardware
side. It was faulty hardware, but since the application 'worked in
DOS' I got no satisfaction from the outlet. Only when I talked
with person who handled the region of all the Southern states did
it get fixed and fast.
That same company sent out a repair person [the HW was under
contract support] who replaced a motherboard that was failing to
see all the RAM. That person was a buy the book in that you put in
this piece, try it, put in next piece.
I was told that she was on the phone talking with the remote
support people who were talking her through the replacements, and
she turned it on, and then she told the support person, now there
is smoke coming out of the machine, and the far side person had to
say 'turn it off'.
Next day there was yet another motherboard. So instability is
often caused by people who know what steps to take and do them
blinding without knowing what they are doing - and that applies to
software, hardware, and all points in between.
I've had SCO machine stay up for 1.5 years at the most. Got 770+
days out of a BSD, and my longest Linux uptime have never been more
than 6 months BUT that was NOT a Linux problem - it was a long
power failure that was longer than UPS. But that system didn't
come back to life because of the way xinetd would start everyting
up, but when it found ONE program that did not start correctly, it
turned off all other programs that it had started. That is my one
major complaint about Linux so far - but now I know what the
problem is in that system I'll work around it.
And knowing how to work around a problem so it won't happen again
is one way to make it stable. But given and equal level of
expertise almost all *n*x systems are stable, at least from my POV.
....
>>>So we are to believe that you honestly tell your customers about the
>>>pros and cons of Linux vs. SCO yet you also state that:
>>>>> >.... I refuse to allow linux into the professional
>>>>> >work environments I'm responsible for?
>>>These two statements cannot be reconciled. You have shown your
>>>true attitude here: you have some irrational bias against Linux
>>>and you are most likely doing your customers a disservice when
>>>you advise them.
>>All Brian said was that he refuse to allow it in the environments
>>for which he's responsible. That's not much difference from
>>someone who has a bias saying "I will never own a Ford" [or
>>whatever the object of derision will be].
>It's OK to say "I will never own a FORD" -- it's OK to make an
>irrational decision that only affects yourself.
>It's OK for a Ford dealer to only sell Fords. Everyone understands that
>a Ford dealer will not give impartial advice.
>What is not OK is to say that you will act in the best interests
>of your client yet rule out the possiblity that one particular OS
>could ever be used. His statement is equivalent to saying that
>he can never envisage a situation where Linux would be the best
>alternative.
And he did say this about 'systems for which he is responsible'.
So if Brian does not like Linux for any reason, real or imagined,
and he has to support them, then he has two choices. The most
logical one is 'don't install linux'. The other alternative is
"fire me and find someone else".
That sometimes is a good alternative. Years ago on a Xenix system
with a non-standard SCSI controller I was doing programming and
support on a system that others installed and spec'd. One day it
failed and the fsck destroyed the system. Since it was a 'custom'
way to build the boot disks, which had to be build on a non-SCSI
machine, I could not fix it.
He called is HW vendor who said "what did he do to my machine". And
I told him what the problem was. I have never seen a person so mad
or red in the face he was not shouting. But I do recall close to
what I said. "They are saying one thing, and I'm telling you the
exact opposite, and you have no way to know which is right. The
only solution I see is to fire us both and find someone else".
He wound up callling SCO, they put him in touch with an authorized
support house, he came out, looked at things. And two week later I
was back installing a correct version for the HW. Trying to
salvage an untenalbe position by not admiting you can't handle it,
or making some excuse is far worse than walking away. In the end
the only thing that has validity is 'will the customers system do
the work they need done in the best way for them'.
>>He's not doing them a dis-service if they want him to work for them
>>and he's not comfortable working in that environment.
>If the client understands that there could be a situation where the
>client will not get the best alternative because Brian will never
>recommend Linux, then I agree. But I rather doubt that Brian has made
>this particular point clear to his customers.
I have no idea of what Brian may have done or not done. But I have
seen posts by his on this and other places where he does know what
he is doing.
>I suspect that Brian rationalizes it by thinking that Linux
>COULD NEVER BE the best alternative. I doubt many people here
>would agree with such a sweeping claim.
What neither you or I know is exactly what customer base Brian is
dealing with. You can come up with a scenario where certain OSes
should never be considered for a particular type of
operaton/business - and the reasons could vary. So if Brian won't
recommend Linux for any of his clients, and to judge that statement
valid, would we not have to know exactly what his client base
consists of?
>>If the customer wants Linux and he doesn't care to support it,
>>he should point them to someone who does.
>Does he do this? I believe he does his best to force an SCO
>alternative.
Don't most people who sell/support something do the best to keep
the customer with what they sell/support. Life is not like the
Santa in Miracle on 34th Street where Santa says "we don't have
this but Macy's does".
Any one pushing Linux where SCO would be better would fall into
the same category. There are vertical aps that are supported best,
and sometimes only, on one OS variant. I had one client whose SW
vendor was pushing them very hard to move from SCO into an AIX
system on an RS6000. That vendor had recommended AIX for all
systems over 16 users, and kept SCO for 16 users and under. This
client [ who is now self-supporting with a full-time computer
person on staff ] was running 122 uses the last time I saw it.
And a price far cheaper than the only other OS the vendord
supported, AIX.
>>And there are others who have migrated to other platforms other
>>than Linux - though Linux is the 'hot' brand now.
>And I don't have a problem with migrating to other OSes. I do
>have a problem with people claiming that Linux can never be the
>best solution.
There is no one OS that is best for everything. But Brian never
said that Linux can't be the best solution except for HIS current
customers.
>Brian's alternative appears to be SCO. Yet, SCO has reduced it's
>development staff and bet the company on lawsuits that appear to be
>very unlikely to pay off. What is Brian going to tell his customers if
>they run into a bug and there is no support because the SCO is a
>bankrupt shell?
Considering the first SCO systems I did work on date from about
1984 and they were almost all just one major application - I've not
seen that many bugs. I don't recall any that were show stoppers
that didn't involve a 3rd party driver and the 3rd party got those
fixed.
I would wait until all the legal disputes are finshed to see what
the final outcome is before speculating on a future events which
may or may not happen.
>IF< SCO were to go bankrupt, that would be the time to 1)
recommend someone else or 2) convince the customer that you can fix
any problems that may crop up or 3) something completely different.
I think Chicken Little was wrong.
Bill
-- Bill Vermillion - bv @ wjv . com
- Next message: Bill Vermillion: "Re: No longer supporting Unixware / Open Server"
- Previous message: J. L. Schilling: "Re: No longer supporting Unixware / Open Server"
- In reply to: Joe Dunning: "Re: No longer supporting Unixware / Open Server"
- Next in thread: Brian K. White: "Re: No longer supporting Unixware / Open Server"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
