Cranky Guru's in NG
zerocool
to help diagnose problems. Do you people that are guru's and well respected
in this newsgroup expect to get more people involved in the UNIX community
by flaming them for their ignorance? I come from a Windows95,98,NT(nice try)
background and know the pain we all feel when a newbie asks a question
without giving all the details. The more people we scare away from the UNIX
community by doing this the more are going to jump on the Windows 2000
point-and-click bandwagon.
At least they are trying to get into UNIX, which, for some of us,
has been a real effort. A little slack for the newbies would be nice, not
all of us cut our teeth on UNIX..................
Art Armfield
SCO CUSA
Tony Lawrence
>
> Ok, OK, so some newbies are not giving anyone all the information they need
> to help diagnose problems. Do you people that are guru's and well respected
> in this newsgroup expect to get more people involved in the UNIX community
> by flaming them for their ignorance?
The only recent flaming around here was of somebody who
basically said "It doesn't install" which is a dumb and
useless post no matter what the OS is. And none of that was
even flames- just notes that the post was useless and
couldn't possibly get a useful answer.
> I come from a Windows95,98,NT(nice try)
> background and know the pain we all feel when a newbie asks a question
> without giving all the details. The more people we scare away from the UNIX
> community by doing this the more are going to jump on the Windows 2000
> point-and-click bandwagon.
Most of us bend over backwards to accommodate newbies. But
even newbies have to exhibit a little common sense. You
don't call up your mechanic and say "My car doesn't work
right". You at least get somewhere close to the real
problem- it doesn't start, it's stuck in first gear, it's
making a strange noise, whatever.
The person referred to said (paraphrased) "I've followed the
directions in the manual but it doesn't work". The manual
is more than 500 pages- which directions do you suppose they
meant? If they had offered more than "it doesn't work"
that might have helped- like "it never recognizes my IDE
CDROM", or "it says there's no root drive controller" or
SOMETHING.
Look, I agree that sometimes people can be too curt and
impatient. I've complained myself that replies of the "man
sendmail" type are often useless- but there are exceptions
(for example, people often ask questions about ftp that are
covered in the man page for "ftpaccess"). But people who
effectively haven't asked a question that can be even
guessed at need to be told so- cripes, is it so awful to
make a legitimate complaint about a COMPLETE lack of
information?
By the way, be careful of slinging around that "guru"
title. There are a few real gurus and one or two of them
post here infrequently, but most of the people who think
they deserve that title do not, and those of us who know
damn well that we aren't even close get uncomfortable when
that kind of noise is made around us.
--
Tony Lawrence (to...@aplawrence.com)
SCO/Linux articles, help, book reviews, tests,
job listings and more : http://www.pcunix.com
Wolf Grossi
helpful.
Wolf
Tony Lawrence
>
> In my opinion, this ng's responses are 99.9% very professional, kind &
> helpful.
Oh, bite me.
No, just kidding :-)
Certainly we all get cranky now and then, but most of us are
here to help, not to show off our supposed brilliance. Way
back when I was even more of a fumbling dolt than I am
today, I learned a lot from this very newsgroup (well,
technically the name has changed..) and because of that
knowledge I';ve enjoyed a pretty good standard of living for
quite a while now. Some of the people who were active
posters then are gone now, some remain, but the questions
still keep coming and I feel an obligation to pay back some
of my debt when I can.
And again, I'm no guru, and neither are most of the nice
folks who post regularly and try to help. Some of us can
lay claim to being "professional", and most of us have
learned a trick or two in our years of experience, but to my
mind, someone like Ken Thompson is a guru and we don't get
the benefit of people of that stature around here very
often..
Jeff Liebermann
wrote:
>Ok, OK, so some newbies are not giving anyone all the information they need
>to help diagnose problems. Do you people that are guru's and well respected
>in this newsgroup expect to get more people involved in the UNIX community
>by flaming them for their ignorance?
Interesting theory. By being polite, nice, considerate, and
diplomatic, I'm suppose to be selling Unix to the GUM (great unwashed
masses). It doesn't work that way.
1. Unix gurus are EXPECTED to be arrogant, obnoxious, short tempered,
and egotistical. Without these distinguishing characteristics, they
would not be a Unix guru, just another burned out tech support
survivor. I can get the most polite wrong answers from a variety of
vendors support organizations, but I prefer correct information.
2. There's a reason that the pharmacutical industry add foul tasting
flavours to their pills. People don't believe that effective pills
should taste good. I tried being nice and diplomatic and found that
nobody was following my advice or solutions. When I switched to being
mean and nasty, my credibility was instantly assumed, and people
actually started following my suggestions and learning as they went.
3. There is no way that my answers to questions have any effect on
SCO Unix sales. 99.9999% of the questions in this newsgroup come from
people that already own and operate SCO Unix. However, if SCO wants
to retain my services in some promotional capacity, I would happily
accept any stipends, bribes, pay-offs, etc.
4. I do this for free. That gives me the license to act in most any
variation on civilized behavior. I've tried just about every method
imaginable to get people with questions to supply the exact version
number of their product. Asking for updates installed required
begging. I've done this for a long time and found that being mean and
nasty to anyone that doesn't supply a version number works very well,
as repeated postings usually include that vital piece of information.
For a while, I simply ignoring any questions that didn't include a
version number. When I discoved that only about 30% had version
numbers, I gave up on that incentive plan.
5. Doing this for free does not automatically assume that my time is
worthless. Figure on $75/hr. It takes me about 15-30 minutes to
research and fabricate a suitable reply. If someone with a question
doesn't bother to supply even the minumum amount of information
necessary to supply a useable answer, I don't see why I should burn my
time attempting to do guesswork where 5 seconds of reading the screen
would have supplied the necessary information. However, I think the
phenomenon is univeral. A friend at the parts counter at a local
automotive parts dealer tells me that many customers don't realize
that the parts guy needs to know the make, model and year to find the
parts.
>I come from a Windows95,98,NT(nice try)
>background and know the pain we all feel when a newbie asks a question
>without giving all the details. The more people we scare away from the UNIX
>community by doing this the more are going to jump on the Windows 2000
>point-and-click bandwagon.
Do you have any statistics that demonstrate this contention? People
switch to Windoze for a variety of reasons. Lack of wiz-bang
applications for Unix is the main problem. Somewhere near the bottom
of the list must be obnoxious gurus in newsgroups.
> At least they are trying to get into UNIX, which, for some of us,
>has been a real effort. A little slack for the newbies would be nice, not
>all of us cut our teeth on UNIX..................
It works both ways. Make my life easier to answer your questions by
doing some light homework first, and supplying enough information to
answer the question. Some clue as to what you're trying to accomplish
is also helpful.
I have a better idea. For every (free) answer you get, try answering
someone elses question. There are plenty of Windoze related questions
in this group. Some of the basic stuff only requires a little RTFM or
pointers to FAQ's, web sites, applications, Skunkware, etc. I
actually practiced this in several newsgroups where I know very little
about the topic. Want someone to spend 15-30 minutes on your problem?
Fine, then spend the same time answering someone elses.
Meanwhile, bah-humbug.
--
Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
(831)421-6491 pgr (831)426-1240 fax (831)336-2558 home
http://www.cruzio.com/~jeffl WB6SSY
je...@comix.santa-cruz.ca.us je...@cruzio.com
Bill Vermillion
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>On Sun, 23 Apr 2000 09:20:22 -0400, "zerocool" <zero...@acidburn.com>
>wrote:
>>Ok, OK, so some newbies are not giving anyone all the information
>>they need to help diagnose problems. Do you people that are
>>guru's and well respected in this newsgroup expect to get more
>>people involved in the UNIX community by flaming them for their
>>ignorance?
>1. Unix gurus are EXPECTED to be arrogant, obnoxious, short tempered,
>and egotistical. Without these distinguishing characteristics, they
>would not be a Unix guru, just another burned out tech support
>survivor. I can get the most polite wrong answers from a variety of
>vendors support organizations, but I prefer correct information.
Actually the few true Gurus I have met don't think they are Gurus
and they only time they get short tempered is with those who won't
help themselves. A couple were mildy arrogant. Maybe I've just
met a better class of gurus' :-) Most however did fit the mold
that goes with the old Southern saying "My ma didn't tolerate no
fools". You didn't get a tongue-lashing if you asked a question
without thinking. It was more like being sliced into deli-thin
portions by a razor sharp rapier. You typically only asked ONE
question without thinking about it first.
I have a friend I'd consider a guru - though he doesn't consider
himself one - and his new employer needed to have him certified. Ten
months from the first test through passing the CCIE lab qualifies
him both as a Unix and network guru - at least in my book.
>2. > I tried being nice and diplomatic and found that
>nobody was following my advice or solutions.
Why should you care if they follow your advice? You gave the
advice it's up to them to make the choice of whether to follow it
or not. If they don't it's their loss.
>5. ...
> It takes me about 15-30 minutes to research and fabricate a
>suitable reply.
It's not the researched ones I'd worry about - it's the ones you
fabricate. :-)
>Meanwhile, bah-humbug.
Nah. Not "Bah-humbug". That is associated with Scrooge and
Christmas. Today is Easter. So it's "That cwazy wabbit"
:-)
Bill
--
Bill Vermillion bv @ wjv.com
Bill Vermillion
Tony Lawrence <to...@aplawrence.com> wrote:
>Wolf Grossi wrote:
>>
>> In my opinion, this ng's responses are 99.9% very professional, kind &
>> helpful.
>Oh, bite me.
>No, just kidding :-)
>Certainly we all get cranky now and then, but most of us are
>here to help, not to show off our supposed brilliance. Way
>back when I was even more of a fumbling dolt than I am
>today, I learned a lot from this very newsgroup (well,
>technically the name has changed..) and because of that
>knowledge I';ve enjoyed a pretty good standard of living for
>quite a while now. Some of the people who were active
>posters then are gone now, some remain, but the questions
>still keep coming and I feel an obligation to pay back some
>of my debt when I can.
In the group of computer nuts I used to hang around with, it was
called "pay forward". Since we can't 'pay back' those who taught
us as they had a head start and we still learn from them, we helped
the newer people, and thus 'pay forward'. It still seems to work
well.
Jim Richardson
"Wolf Grossi" <wg_...@magro-soft.com> wrote:
> In my opinion, this ng's responses are 99.9% very professional, kind &
> helpful.
>
>
We've just been lucky enough to have made 99.9% more mistakes than some
others. :)
Seriously, nearly every repose to a request for help in this newsgroup
comes from one of us who has made the same mistake or faced the same
problem that the poster is asking about. It's not that we know more,
we've just made more mistakes.
--
Jim Richardson
I like NT because it constantly reminds me of my daughter.
"Honest Daddy, I wasn't doing anything and it just broke."
Sent via Deja.com http://www.deja.com/
Before you buy.
Tazman37
OK, OK. I had it coming I guess. Especially from Tony whom I respect
greatly. His web page alone has bailed my incompetant butt out at least a
hundred times. Just trying to make a point and I guess it got a little out
of hand.
Loved your article on "unhappy modems" BTW Tony. I have one
driving me crazy right now.
I do apologize if I stepped on any toes, please feel free to personally
flame me at my email address which I will post here and for which I stand
corrected:
Art Armfield
SCO CUSA
zerocool <zero...@acidburn.com> wrote in message
news:3902f986$0$97502$2486...@news.freeway.net...
> Ok, OK, so some newbies are not giving anyone all the information they
need
> to help diagnose problems. Do you people that are guru's and well
respected
> in this newsgroup expect to get more people involved in the UNIX community
> by flaming them for their ignorance? I come from a Windows95,98,NT(nice
try)
> background and know the pain we all feel when a newbie asks a question
> without giving all the details. The more people we scare away from the
UNIX
> community by doing this the more are going to jump on the Windows 2000
> point-and-click bandwagon.
> At least they are trying to get into UNIX, which, for some of us,
> has been a real effort. A little slack for the newbies would be nice, not
> all of us cut our teeth on UNIX..................
>
>
> Art Armfield
> SCO CUSA
>
>
>
>
Bill Vermillion
Tazman37 <zero...@acidburn.com> wrote:
>OK, OK. I had it coming I guess. Especially from Tony whom I
>respect greatly. His web page alone has bailed my incompetant butt
>out at least a hundred times. Just trying to make a point and I
>guess it got a little out of hand.
> Loved your article on "unhappy modems" BTW Tony. I have one
>driving me crazy right now.
> I do apologize if I stepped on any toes, please feel free to
>personally flame me at my email address which I will post here and
>for which I stand corrected:
What else are we going to use toes for if not to step on.
If anyone is afraid of getting flamed or having barbs thrown in
their direction - then they shouldn't be posting to the 'net.
We've all helped create/perpetuate this maelstrom and that brings
to mind a quote from a famous fictional Southern entity of years
gone by - Pogo 'Possum. "I've seen the enemy, and they is us!".
Bill
>armf...@freeway.net
>
>Art Armfield
>SCO CUSA
>
>
>zerocool <zero...@acidburn.com> wrote in message
>news:3902f986$0$97502$2486...@news.freeway.net...
>> Ok, OK, so some newbies are not giving anyone all the information they
>need
>> to help diagnose problems. Do you people that are guru's and well
>respected
>> in this newsgroup expect to get more people involved in the UNIX community
>> by flaming them for their ignorance? I come from a Windows95,98,NT(nice
>try)
>> background and know the pain we all feel when a newbie asks a question
>> without giving all the details. The more people we scare away from the
>UNIX
>> community by doing this the more are going to jump on the Windows 2000
>> point-and-click bandwagon.
>> At least they are trying to get into UNIX, which, for some of us,
>> has been a real effort. A little slack for the newbies would be nice, not
>> all of us cut our teeth on UNIX..................
>>
>>
>> Art Armfield
>> SCO CUSA
>>
>>
>>
>>
>
>
Tony Lawrence
>
> OK, OK. I had it coming I guess. Especially from Tony whom I respect
> greatly. His web page alone has bailed my incompetant butt out at least a
> hundred times. Just trying to make a point and I guess it got a little out
> of hand.
> Loved your article on "unhappy modems" BTW Tony. I have one
> driving me crazy right now.
That's an oldie :-) I had almost forgotten it- I wrote that
a long, long time ago.
Either modems have gotten better or I've gotten smarter-
probably a little of both- but I haven't had a modem tick me
off in quite a few yesrs now..
Jeff Liebermann
Vermillion) wrote:
>Actually the few true Gurus I have met don't think they are Gurus
>and they only time they get short tempered is with those who won't
>help themselves. A couple were mildy arrogant. Maybe I've just
>met a better class of gurus' :-)
Most gurus are fairly normal in person. However, just try to get into
an arguments with one. Instant arrogance. They weren't always like
that. It's just a side effect of knowing too much.
>Why should you care if they follow your advice?
I keep score. So far, my batting average for guesswork is about 50%.
Others do much better. I can't tell if I'm right or wrong unless they
at least follow my suggestions and report back the results. The ones
that *REALLY* piss me off are a full repost of my wisdom with "Thanks
Much" tacked on top. No way to know if I was right or wrong. No clue
if I was summarily ignored, or if my guesswork was religiously
followed. No clue if they're going to build a statue in my honor, or
an effigy.
>It's not the researched ones I'd worry about - it's the ones you
>fabricate. :-)
It's fairly scarey when some people can't tell the difference between
my humor and reality. Sometimes, it scares me too. I will confess to
embellishing some of my more creative fabrications with a liberal dose
of truth, but that's the exeption rather than the rule.
>Nah. Not "Bah-humbug". That is associated with Scrooge and
>Christmas. Today is Easter. So it's "That cwazy wabbit"
You celebrate Easter with the Playboy bunny? Sounds like fun.
That might explain why Intel trademarked and registered the domain
"bunnypeople".
Danny Aldham
zerocool <zero...@acidburn.com> wrote:
> Ok, OK, so some newbies are not giving anyone all the information they need
> to help diagnose problems. Do you people that are guru's and well respected
> in this newsgroup expect to get more people involved in the UNIX community
> by flaming them for their ignorance? I come from a Windows95,98,NT(nice try)
> background and know the pain we all feel when a newbie asks a question
> without giving all the details. The more people we scare away from the UNIX
> community by doing this the more are going to jump on the Windows 2000
> point-and-click bandwagon.
He-he-he. Take a stroll over to the perl or cgi groups. They make it
look like kindergarten (ie.holding hands & wiping noses) over here.
--
Danny Aldham Providing Certified Internetworking Solutions to Business
www.postino.com E-Mail, Web Servers, Mail Lists, Web Databases, SQL & Perl
Chad Lemmen
"zerocool" <zero...@acidburn.com> wrote:
> Ok, OK, so some newbies are not giving anyone all the information they
need
> to help diagnose problems. Do you people that are guru's and well
respected
> in this newsgroup expect to get more people involved in the UNIX
community
> by flaming them for their ignorance? I come from a
Windows95,98,NT(nice try)
> background and know the pain we all feel when a newbie asks a question
> without giving all the details. The more people we scare away from the
UNIX
> community by doing this the more are going to jump on the Windows 2000
> point-and-click bandwagon.
us,
> has been a real effort. A little slack for the newbies would be nice,
not
> all of us cut our teeth on UNIX..................
>
> Art Armfield
> SCO CUSA
I don't agree with you. I have to say that this newsgroup is the best
of any newsgroup I have ever posted to. It has some of the most
knowledgeable UNIX people in it that I have ever come across. This
newsgroup is a major reason why I have embraced UNIX and come to love it
over the past 5 years. I have never posted a question here that has not
been answered. Most of my problems have been solved either the same day
or the next day. Thats better service than most software vendors I've
dealt with offer. Best of all its free. Nobody makes any money spending
time answering these posts, which makes me appreciate the time they take
all the more. Most everyone in here gives very complete and easy to
follow instructions. Take a look at some of Jeff Liebermann's or Tony
Lawrence's posts. There are many others sorry for not mentioning
everyone, but those two come to mind. Jeff especially has really helped
me out a lot over the years. Also you mentioned being flamed for not
including enough info in the post. I wouldn't call it being flamed. It
just makes me try to make better posts by being more complete. These
guys know their stuff, but you have to do your part by giving all
relevant info or its hard for them to diagnose the problem. I've tried
posting to some of the Windows newgroups, but never got any intelligent
answers in there if any answers at all. In the Windows newsgroups there
are a lot of posts asking questions, but few of them get answered. In
here I think most everything gets answered. Keep up the good work guys
I for one wish more newsgroups were like this one.
Tony Lawrence
> I've tried
> posting to some of the Windows newgroups, but never got any intelligent
> answers in there if any answers at all.
It's not their fault. There are no intelligent answers for
Windows problems.
Actually, I'm at least half serious.
I was working with an NT administrator yesterday on a Unix
problem. What had happened was that we had installed a few
things: a RealWorld upgrade, Visionfs 3.0, and Powerchute
software. After all that, people couldn't print any more.
It took me a little bit of work to track this down, because
the NT guys trouble shooting skills were poor- here's what I
was initially told:
- none of the printers work, including a serial dot matrix.
- the printers work fine from the command line, but not from
RealWorld
I first had him check /usr/spool/lp/tmp and
/usr/spool/lp/logs/requests and learned
that nothing was backing up in the spool directory, but the
log files showed that the RealWorld jobs had been printed.
My first suspicion was RealWorld configuration, but I
eliminated that pretty quickly. It then turned out that all
of his command line tests were done as root, and all of his
RealWorld tests were done as an ordinary user. I had him
try Realwoorld as root, and it worked. So the reality was
now that printing worked for root, but not for ordinary
users. That, of course, made my suspicious of perms on lp,
but those were OK. I wasn't thinking of anything related
to hpnp because the serial printer didn't work either.
Ah, but it turned out that that wasn't quite true either-
the serial printer had been disabled and also was not even
accepting requests. In fact, upon close reexamination,
there were no entries for the serial printer in the logs.
So I ran /usr/lib/accept and enabled it and now I knew that
the problem really related to the HP network printers only.
But it couldn't be a network problem because root could
print. Therefore it had to be related to something the hp
script does.
If you look at the HPNP interface script in
/usr/lib/hpnp/hpnp.model, you see that it also writes its
own log file in /tmp, and only removes it if the print is
successful- so I looked in /tmp, but there were no hpnp log
files. I then checked perms on /tmp, and saw why- it had no
write permission. How did that happen? I doubtedcit was
RW, and it sure couldn't have been Visionfs, so that left
APC Powerchute, and tha indeed was the source.
The actual problem was that the Powerchute install script
stupidly changes perms on the directory it is installed from
(why?). It had changed /tmp and caused the failure.
Two comments the NT admin made bothered me. The first was
"I wouldn't have known where to look". OK, that's fair
enough, but there are two unrelated things to point out in
that regard. The first is that at least you CAN look (you
can't look at the innards of NT's printing system). The
second is that his trouble shooting wasn't very good- using
root to test command line printing and an ordinary user to
test RW clouded the real issue, and not noticing that the
serial printer jobs did NOT appear in the lp logs also
masked the real issue. But, I can't complain too much- I
earn a good living because my trouble shooting skills ARE
good, and I can't expect everyone to have such skills- if
everyone did, I'd be broke!
It was the second comment that really bothered me: "It's
this kind of stuff that drives people to NT- I've got to
find an NT accounting system".
Cripes!!!
Like Powerchute or some other vendor couldn't screw up your
life on NT? And if they do, just how in hell are you going
to fix it? Chances are, you won't- not without help from
them, assuming they even know what happened. Most NT
"fixes" I've seen have been reinstalls or gross restoration
of the registry- certainly the skills of a typical
administrator do not allow for very much poking around to
identify the real source of a problem. If something like
that had happened on NT, I bet it wouldn't have been fixed
for days, and probably would have required direct assistance
from HP.
Oh well. I'm sure that company will dump its "awful Unix"
box as soon as it can. Such a shame, isn't it?
Ben Rosenthal
It's just plain laziness. He does not want to invest the time to learn the
basics so he can solve this type of problem himself. He does not realize
that the time spent learning UNIX pays dividends in flexibility and
reliability. Administering UNIX allows one to be responsible for the success
of the system, but this is too much responsibility for some people. If you
can not configure or troubleshoot the OS operation then you are not
responsible for any problems. People who like UNIX like the ability to
control the OS.
>
> Like Powerchute or some other vendor couldn't screw up your
> life on NT? And if they do, just how in hell are you going
> to fix it? Chances are, you won't- not without help from
> them, assuming they even know what happened. Most NT
> "fixes" I've seen have been reinstalls or gross restoration
> of the registry- certainly the skills of a typical
> administrator do not allow for very much poking around to
> identify the real source of a problem. If something like
> that had happened on NT, I bet it wouldn't have been fixed
> for days, and probably would have required direct assistance
> from HP.
>
> Oh well. I'm sure that company will dump its "awful Unix"
> box as soon as it can. Such a shame, isn't it?
Yes, but I have a customer that was a staunch NT shop and now really respects
the SCO uptime.
>
>
> --
> Tony Lawrence (to...@aplawrence.com)
> SCO/Linux articles, help, book reviews, tests,
> job listings and more : http://www.pcunix.com
Ben Rosenthal
BRC
Optical Retail POS Solutions
Bill Vermillion
<to...@aplawrence.com> wrote:
>Chad Lemmen wrote:
>
>> I've tried posting to some of the Windows newgroups, but never
>> got any intelligent answers in there if any answers at all.
>It's not their fault. There are no intelligent answers for
>Windows problems.
>Actually, I'm at least half serious.
Oh. I though you were half joking.
>I was working with an NT administrator yesterday on a Unix
>problem. ...
>The actual problem was that the Powerchute install script
>stupidly changes perms on the directory it is installed from
>(why?). It had changed /tmp and caused the failure.
There are many programs that require root installs. However there
are some programs which do things in the proper manner. Many of
these require a short root install which creates a user who owns
the application, and perhaps a script in that directory. Then you
log out and back in as that user and install the application. That
gets rid of at least part of the problem. Vendors whose installs
break parts of the OS should be made aware of that fact. If they
pay not attention, then it should be posted on the newsgroup and
posted in the FAQs with warnings on the problems and how to correct
them.
>Two comments the NT admin made bothered me. The first was
>"I wouldn't have known where to look".
I semi-maintain one machine in an NT environment. The amdinstrator
there used the standard MS routines when the modems would fail to
respond, or if something else wasn't working as it should. They'd
announce to everyone over the phone/pa to get out of the Unix system
an then reboot it. Heaven help the person who was in the restroom
or out of their office when this occured.
I got her to run a ps -ef when she had a problem before the reboot,
and identified one problem and fixed that. So much of the MS
philophys just says 'reboot' to see if it fixes it instead of
trying to find the cause.
The typical response is 'well I can get things running more quickly
by just rebooting'. That dosn't take into consideration how much
real producition time is loss when you total the sum of the
reboots. It's a user/adminstration mind-set that needs to be
overcome.
>OK, that's fair
>enough, but there are two unrelated things to point out in
>that regard. The first is that at least you CAN look (you
>can't look at the innards of NT's printing system). The
>second is that his trouble shooting wasn't very good- using
>root to test command line printing and an ordinary user to
>test RW clouded the real issue, ...
But we all have notice that some people use root routinely.
We've notice more than one post from 'root' in these newsgroups.
Last week's problem on another system was caused by the adminstrator
who is the only person who routinely logs in as root and runs all
the applications as root - while all others have their own login.
>It was the second comment that really bothered me: "It's
>this kind of stuff that drives people to NT- I've got to
>find an NT accounting system".
>Cripes!!!
Let's hope he's not representative of all NT admins.
> Most NT "fixes" I've seen have been reinstalls or gross
>restoration of the registry- certainly the skills of a typical
>administrator do not allow for very much poking around to identify
>the real source of a problem. If something like that had happened
>on NT, I bet it wouldn't have been fixed for days, and probably
>would have required direct assistance from HP.
And I've seen posts in other groups where the users were influenced
by the MS philophy because the people with the *ix problem had
indicated that they had to reinstall the system to make it work.
The three R's of MS strike again, reboot, reinstall, replace.
>Oh well. I'm sure that company will dump its "awful Unix"
>box as soon as it can. Such a shame, isn't it?
The shame is that it's happening on small/medium systems where the
company is of such a size where the NT overhead sometimes adds more
of a burden. Once a company realizes that it's not either Unix OR NT
and that both can co-exists do we get over that hurdle. The problem
is getting them to realize it.
Jeff Liebermann
>It's not their fault. There are no intelligent answers for
>Windows problems.
Ah, but Windoze has wizards. Who needs intelligent answers when the wizard
can walk you though a re-install?
>Actually, I'm at least half serious.
You're wrong. I've been answering questions long enough to know that there
are other factors involved. The big one is whether the administrator is
capable of following my instructions and utilizing my advice. I can deliver
the answer to a problem, and some people that have relied on wizards will
not be able to figure out how to use the information. They would require a
step by step set of instructions of precisely what to inscribe on the
command line, and what the check boxes should look like, and only then can
they fix a problem.
Windoze admins tend toward the heavy-hammer approach to fixing problem. The
tradional method of fixing a broken Windoze box is to reinstall everything.
I've seen the trend in this newsgroup for people asking how to re-install
Unix to fix some minor problem. In a way, I don't blame them as a
meticulous approach to troubleshooting is very frustrating with Windoze.
The lack of suitable system diagnostics and version control makes systematic
troubleshooting almost impossible. Therefore, the only fix that works
almost all the time, and for everything, is to reinstall Windoze. Do it
often enough and it's actually easier than figuring out what is wrong.
Another difference is that a Unix admin is expected to know how things work.
A Windoze admin is expected to only know how to make things work. The
distinction is that one cannot troubleshoot problems without understanding
the underlying technology. I'm not referring to kernel or OS internals, but
the basic functionality of the operating system. Unix troubleshooting
emphasizes knowing how the parts talk to each other and how they function
together. Windoze troubleshooting reserves such knowledge for developers.
I firmly believe that I can fix anything if I know exactly how it operates.
>It was the second comment that really bothered me: "It's
>this kind of stuff that drives people to NT- I've got to
>find an NT accounting system".
Can you blame him? With NT, all he needs to do is say "It's busted. Time
for a reinstall". With Unix, he actually has to find and fix the problem.
Which is easier? Before you go on about which is better, realize that the
real purpose of system administrators is to keep the phone from ringing. If
reinstalling NT every few months is an acceptable alternative to spending
considerable more time learning how to troubleshoot Unix, then he's
accomplished his primary function.
>Like Powerchute or some other vendor couldn't screw up your
>life on NT?
Nice troubleshooting job. You were very logical and systematic in finding
and fixing the problem. I'm not. I would have run:
custom -V
custom -v strict
and let it run. It should have reported the tweaked directory permissions.
I also have tripwire installed to check for changes on key files and
directories on some systems.
Methinks you've seen my previous evaluation of Powerchute software and my
suggestion of the precise recycling method that APC can do with it. They
want encrypted passwords in /etc/passwd instead of /etc/shadow as they can't
seem to figure out how to deal with authentication.
>Most NT
>"fixes" I've seen have been reinstalls or gross restoration
>of the registry- certainly the skills of a typical
>administrator do not allow for very much poking around to
>identify the real source of a problem.
Exactly. That's why SCO's IPU (In Place Upgrade) and ability to reinstall
on top of an existing installation both suck. Reinstallation is not an
option with SCO Unix as it would destroy rather than fix the system. Even
reinstallation of components is problematical and difficult (with rollbacks
for added complexity). However, I kinda like it that way. If SCO actually
put some time and effort into wizardizing the installs and making the IPU
more bulletproof, you would have your lazy admins, and clueless beginners
doing reinstalls instead of troubleshooting. Big hammer or little hammer?
Doesn't matter, whatever works is best.
>If something like
>that had happened on NT, I bet it wouldn't have been fixed
>for days, and probably would have required direct assistance
>from HP.
HP would have said "Reinstall everything".
Tony Lawrence
>
> On Sat, 29 Apr 2000 13:57:33 GMT, Tony Lawrence <to...@aplawrence.com> wrote:
>
> >It's not their fault. There are no intelligent answers for
> >Windows problems.
>
> Ah, but Windoze has wizards. Who needs intelligent answers when the wizard
> can walk you though a re-install?
>
> >Actually, I'm at least half serious.
>
> You're wrong.
Which half is wrong- the serious half or the not serious
half?
>I've been answering questions long enough to know that there
> are other factors involved. The big one is whether the administrator is
> capable of following my instructions and utilizing my advice.
You have to identify the problem first- when the
administrator doesn't accurately describe the problem, it
takes much longer. And that's the real measure of computer
skill, IMNSHO: being able to accurately define what the
present situation really is.
>I can deliver
> the answer to a problem, and some people that have relied on wizards will
> not be able to figure out how to use the information. They would require a
> step by step set of instructions of precisely what to inscribe on the
> command line, and what the check boxes should look like, and only then can
> they fix a problem.
I know. It drives me insane. I worked with another
consultant once. The actual client thought this guy was
brilliant- well, heck, he has two (count 'em) doctorates and
half a dozen patents, he must be brilliant. To my mind, he
was the stupidest thing I've come across in years. I was
giving him commands over the telephone, things like "cd
SPACE slash somewhere ENTER ell-ess SPACE minus-ell SPACE
asterixlog". After doing this probably a dozen times, I
forgot to say SPACE after cd so of course he didn't type
it. Didn't have a g.d. CLUE why it hadn't work. So I
briefly explained that you have to type the space, always,
and we continued. A few commands later, I forgot it again.
The jackass folowed my literal instructions again, again did
not understand why it failed, and when I reminded him of the
necessity of the space after cd, he chastised me for not
being accurate! You'd think a guy with all those doctorates
and patents (computer related, no less!) wouldn't need me to
help him anyway..
>
> Windoze admins tend toward the heavy-hammer approach to fixing problem. The
> tradional method of fixing a broken Windoze box is to reinstall everything.
> I've seen the trend in this newsgroup for people asking how to re-install
> Unix to fix some minor problem. In a way, I don't blame them as a
> meticulous approach to troubleshooting is very frustrating with Windoze.
> The lack of suitable system diagnostics and version control makes systematic
> troubleshooting almost impossible. Therefore, the only fix that works
> almost all the time, and for everything, is to reinstall Windoze. Do it
> often enough and it's actually easier than figuring out what is wrong.
Easy for the admin- hell on the users.
>
> Another difference is that a Unix admin is expected to know how things work.
> A Windoze admin is expected to only know how to make things work. The
> distinction is that one cannot troubleshoot problems without understanding
> the underlying technology. I'm not referring to kernel or OS internals, but
> the basic functionality of the operating system. Unix troubleshooting
> emphasizes knowing how the parts talk to each other and how they function
> together. Windoze troubleshooting reserves such knowledge for developers.
> I firmly believe that I can fix anything if I know exactly how it operates.
Exactly. And if you have done some programming, you know
how it HAS to work, or at least what choices the programmer
probably made. That knowledge sometimes amazes people- "How
did you know that would fix it?"- well, 'cause there are
really only a couple of ways the programmer COULD do it, so
even if I can't see his code, I've got a pretty good guess
as to what it must be..
>
> >It was the second comment that really bothered me: "It's
> >this kind of stuff that drives people to NT- I've got to
> >find an NT accounting system".
>
> Can you blame him? With NT, all he needs to do is say "It's busted. Time
> for a reinstall". With Unix, he actually has to find and fix the problem.
> Which is easier? Before you go on about which is better, realize that the
> real purpose of system administrators is to keep the phone from ringing. If
> reinstalling NT every few months is an acceptable alternative to spending
> considerable more time learning how to troubleshoot Unix, then he's
> accomplished his primary function.
>
> >Like Powerchute or some other vendor couldn't screw up your
> >life on NT?
>
> Nice troubleshooting job. You were very logical and systematic in finding
> and fixing the problem. I'm not. I would have run:
> custom -V
> custom -v strict
> and let it run. It should have reported the tweaked directory permissions.
Sure. But that takes too long, and the problem might not
have been perms.
> I also have tripwire installed to check for changes on key files and
> directories on some systems.
Not a bad idea- wouldn't help me much as I'm usually working
on somebody else's systems, but still a good idea.
>
> Methinks you've seen my previous evaluation of Powerchute software and my
> suggestion of the precise recycling method that APC can do with it. They
> want encrypted passwords in /etc/passwd instead of /etc/shadow as they can't
> seem to figure out how to deal with authentication.
I just don't understand what their purpose is in changing
perms. Admittedly I didn't read their installation
instructions, but it's still a dumb thing to do. I'm going
to download that again and see WHY they think they should do
that.
Tony Lawrence
(wrt a Powerchute install screwing with the perms of /tmp)
> I just don't understand what their purpose is in changing
> perms. Admittedly I didn't read their installation
> instructions, but it's still a dumb thing to do. I'm going
> to download that again and see WHY they think they should do
> that.
I downloaded their bundle. The instructions say to untar it
in a "temporary directory" - gosh, I guess they don't mean
/tmp, do they?
The really funny thing about this script is that it is
actually designed to be run off a CDROM- it doesn't even
have a choice for installing from a directory; but since it
has to ask where the CD is mounted, you just give it your
directory instead- that's what the install guide tells you
to do. So doesn't it seem even MORE strange that it would
change perms?
I couldn't duplicate the problem on my system, so I suppose
it's at least possible that Powerchute was not at fault- but
they have code in the script that removes write permission,
so I'm still suspicious..
Raymond N Shwake
>Tony Lawrence wrote:
....
Indeed, what's he going to do when NT's spooler craps out? I've
encountered one such problem using Windows TSE with MetaFrame. One of my
support colleagues received reports that people "couldn't print" to printers
they'd used in the past. I - the UNIX and Network guy - had earlier diagnosed
such behavior. Opening up the Printers windows I found "Add Printer" and
nothing else; *all configured printers were gone*! Your typical NT admin
wouldn't know what to do at this point. As it happens, a smple "net stop
spooler" and "net start spooler" cleared it up.
There's no discussion of such behavior in the NT classes, nor in
online help, nor in the books. NT *can* be fine if your hardware configuration
is well supported with quality drivers, and you keep the demands on the
system to a defined minimum. A load, per se, isn't a problem as long as the
apps are well behaved, don't leak memory, etc. But one shouldn't expect a
stable system if too many apps run on the box, each with their own DLLs, and
so on. When NT *does* go awry, diagnostics can be a bear. That's why the
three most common "resolutions" are: 1) reboot; 2) apply service packs; and
3) reinstall.
Bill Vermillion
Jeff Liebermann <je...@comix.santa-cruz.ca.us> wrote:
>Exactly. That's why SCO's IPU (In Place Upgrade) and ability
>to reinstall on top of an existing installation both suck.
>Reinstallation is not an option with SCO Unix as it would destroy
>rather than fix the system. Even reinstallation of components is
>problematical and difficult (with rollbacks for added complexity).
The SCO system could stand some improvements when compared to other
installs I've used. The SGI IRIX installs are really nice. They
thoroughly check dependancies of other programs, needed libraries,
outdated libraries, etc. And at every step you have the option to
choose to install the piece you wish which will then prompt for
other software, or just skip it.
> However, I kinda like it that way. If SCO actually put some
>time and effort into wizardizing the installs and making the IPU
>more bulletproof, you would have your lazy admins, and clueless
>beginners doing reinstalls instead of troubleshooting.
The above SGI example won't work if you are clueless however.
You do have to know something about what you are installing.
They do have the advantage of working with hardware with few
deviations and since the HW and SW are higher priced than in the
iNTEL based arena they can afford to put more time into their
installs. At least they did before they made some serious
mis-steps in judgement and fell flat - such as their foray into the
NT world on iNTEL based platforms.
>Big hammer or little hammer?
>Doesn't matter, whatever works is best.
>>If something like
>>that had happened on NT, I bet it wouldn't have been fixed
>>for days, and probably would have required direct assistance
>>from HP.
>HP would have said "Reinstall everything".
You also forgon one rule-of-thumb used in probelmatic MS systems.
Before you re-install you reformat everything to 'make sure'.
Jeff Liebermann
>Which half is wrong- the serious half or the not serious
>half?
Both. Users in Windoze newsgroups rarely supply sufficient information to
give a complete answer. I used to hang around news:24hoursupport.helpdesk
and try to answer questions. I gave up simply because very few people were
able to simply transcribble an error message or knew their exact Windoze
version. You're also wrong about there not being any intelligent answers to
Windoze problems. As long as Windoze users are willing to tolerate a
miserable level of reliability, reinstalling the whole mess every few months
is a perfectly intelligent answer that really does work.
>You have to identify the problem first- when the
>administrator doesn't accurately describe the problem, it
>takes much longer. And that's the real measure of computer
>skill, IMNSHO: being able to accurately define what the
>present situation really is.
Correct. However, most NT admins are not being paid to do troubleshooting.
They get paid to keep the phone from ringing and to keep the system alive.
I know one admin whos personal rule-o-thumb is that when he gets forced to
do more than 3 reboots per week, it's time for a total NT4 reinstall. Never
mind the implications of this method of improving uptime. If it's
scheduled, accepted by mis-management, tolerated by the users, and proven to
be functional, it will be done.
>I know. It drives me insane.
Lack of sanity is a pre-requisite for surviving in this business. A sane
person would have run long ago.
>I worked with another
>consultant once. The actual client thought this guy was
>brilliant- well, heck, he has two (count 'em) doctorates and
I'm working with a client that has several remote offices. Each office has
its own guru. The primary contractor has a wall full of diplomas and
awards. When his proposed VPN toplogy was initially proposed, I shot it
full of holes. I scribbled down my predictions and emailed it to the
appropriate parties. It didn't work very well. No problem, as I had a fall
back design ready. Rather than admit that he screwed up, he had every one
of his techs beating their collective heads against the wall trying to make
it work. Someone even dragged in MicroSloth engineering services (don't ask
what that costs). We're 3 months into this fiasco, and he still won't admit
that it won't work. That's the real problem with intelligent people (versus
pragmatic mercinaries like me), they reek of too much ego.
>Easy for the admin- hell on the users.
Nah. Users are easily intimidated. Actually, it's rough on the admins
because the work usually gets done under cover of darkness. Most of my NT4
reinstalls take about 4-6hrs and therefore must be done at night or on
weekends.
>Exactly. And if you have done some programming, you know
>how it HAS to work, or at least what choices the programmer
>probably made.
What? Me program? Surely you jest. I'm a hardware hack. If desperate, I
can scribble some of the worst code known to man or machine. In general, I
just hire a programmer when I need something that actually works. I can
think like a programmer, but no way am I going to perform such an un-natural
act.
>Sure. But that takes too long, and the problem might not
>have been perms.
Sure. However, the long custom running time is a good excuse to get coffee.
>> I also have tripwire installed to check for changes on key files and
>> directories on some systems.
>
>Not a bad idea- wouldn't help me much as I'm usually working
>on somebody else's systems, but still a good idea.
It's a different style of troubleshooting. I try to think like a policeman.
They look for what has changed or what looks out of place. I've seen enough
systems to know how responsive they're suppose to be, how fast, how
organized, how big, etc. When something changes, I notice the change. I
listen to the incoherent complaints, but what I look for are the changes.
That doesn't just mean software. It includes new boxes, lan tweaks,
building construction, coincidental crashes on other machines, irregular
events (end-of-month reports), and other things that are NOT routine. I try
to not thing of it as "finding the problem" but rather "assigning the
blame". That's also why I'm such an SNMP and MRTG fanatic. If I have a
history of what is normal, then I can assign the blame when something
changes.
>I just don't understand what their purpose is in changing
>perms. Admittedly I didn't read their installation
>instructions, but it's still a dumb thing to do. I'm going
>to download that again and see WHY they think they should do
>that.
I've done about 7ea PowerChute Unix installs on various ODT and OSR5 boxes.
I don't recall what I used as an install directory. It *MAY* have been
/tmp. I just followed the instructions. No problems with the /tmp perms.
However, I still think that PowerChute needs lots of work.
Bernd Felsche
>In article <390AEA06...@aplawrence.com>, Tony Lawrence
><to...@aplawrence.com> wrote:
>>Chad Lemmen wrote:
>>> I've tried posting to some of the Windows newgroups, but never
>>> got any intelligent answers in there if any answers at all.
>The typical response is 'well I can get things running more quickly
>by just rebooting'. That dosn't take into consideration how much
>real producition time is loss when you total the sum of the
>reboots. It's a user/adminstration mind-set that needs to be
>overcome.
And you multiply that time by the number of users expecting
service. A total denial-of-service resulting from the reinstall can
be devastating; unless you duplicate your problems to a "mirror"
host.
One major problem is the reluctance (laziness?) amongst
"administrators" is to find the real problem and to fix it so that the
periodic reboots aren't necessary. Having supported some systems which
were only ever rebooted for necessary hardware upgrades, uptime is an
indication of the quality of system support.
You get lazy admins abusing Unix as well. "root" logins are only one
symptom.
--
Real Name: Bernd Felsche
Email: nospam...@perth.DIALix.com.au
http://www.perth.dialix.com.au/~bernie - Private HP
David H. Mabo
Bill Vermillion wrote:
>
> << SNIP >>
> >HP would have said "Reinstall everything".
>
> You also forgon one rule-of-thumb used in probelmatic MS systems.
>
> Before you re-install you reformat everything to 'make sure'.
>
Before you back it up!!!
> --
> Bill Vermillion bv @ wjv.com
--
David H. Mabo, CPCM
Adaptix Corp. - Cincinnati, Ohio
Don Lee
Felsche) wrote:
>
>One major problem is the reluctance (laziness?) amongst
>"administrators" is to find the real problem and to fix it so that the
>periodic reboots aren't necessary. Having supported some systems which
>were only ever rebooted for necessary hardware upgrades, uptime is an
>indication of the quality of system support.
>
>You get lazy admins abusing Unix as well. "root" logins are only one
>symptom.
There is a HUGE difference in the tools available to a UNIX
administrator as compared to an NT administrator. A very typical thing
on an NT system that I have seen is to see from the Task Manager that
you are running out of memory, lets say 256Meg. But if you look at the
individual processes and add up the memory you only see 128 Meg of
memory being used. Even if you start shutting down processes one at a
time you never see the memory usage go down drastically. On an UNIX
system if something like this were to happen (which I have never
seen), I would write a shell script to do a ps -el every now and then
to a file and then look at it every now and then and see if I see
anything strange going on. On NT writing a script requires something
else installed that gives you scripting capability and then you have
to hope you can find some command that gives you what you want thats
not GUI driven or write a program to do what you need. Its just so
much harder to find the problem in NT that I just don't think the
problem is as much lazy admins as it is nearly impossible for NT
admins to figure out.
BTW, I an not an NT or UNIX admin and would never desire to be an NT
admin under any circumstances.
-Don
Bill Campbell
....
>I'm working with a client that has several remote offices. Each office has
>its own guru. The primary contractor has a wall full of diplomas and
>awards. When his proposed VPN toplogy was initially proposed, I shot it
>full of holes. I scribbled down my predictions and emailed it to the
>appropriate parties. It didn't work very well. No problem, as I had a fall
>back design ready. Rather than admit that he screwed up, he had every one
>of his techs beating their collective heads against the wall trying to make
>it work. Someone even dragged in MicroSloth engineering services (don't ask
>what that costs). We're 3 months into this fiasco, and he still won't admit
>that it won't work. That's the real problem with intelligent people (versus
>pragmatic mercinaries like me), they reek of too much ego.
One of my favorite lines came from Anne Fairbairne's ``That Man
Cartwright'' describing the main character's Limousine Liberal wife and her
friends as ``educated beyond their intelligence''. This certainly
describes many of the staff and faculty of our liberal arts colleges and
universities in the U.S. today (have I offended enough people yet :-).
I think that there's a tendency too for people to get to know a lot about
very little, and not to know much at all about things outside their fields.
These people have a tendency to believe ``experts'' ( an Ex is a has-been
and a spurt is a drip under pressure as Dean Atchison said at my high
school graduation -- does that date me) without questioning their
credentials or their real knowledge. This is particularly true if the
expert's opinion supports their preconceived notions and prejudices.
One also has to look at the motives of the expert. Is he going to make a
lot more money by selling an NT solution than he by selling Linux or Unix?
There's a huge industry that depends on Microsoft putting out crappy
software so that they can charge people to keep cleaning up after its
failures. Where would McAfee, Norton, et al be if Windows had real
security and wasn't susceptible to viruses? They would all have to go out
and get Real Jobs(tm) doing things that were productive rather than
cleaning up after Windows.
...
Bill
--
INTERNET: bi...@Celestial.COM Bill Campbell; Celestial Systems, Inc.
UUCP: camco!bill PO Box 820; 6641 E. Mercer Way
FAX: (206) 232-9186 Mercer Island, WA 98040-0820; (206) 236-1676
URL: http://www.celestial.com/
The difference between science and the fuzzy subjects is that science
requires reasoning while those other subjects merely require scholarship.
-- Robert Heinlein