NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/comp/sys/hp/14527

Text File | 1993-01-05 | 5.4 KB | 100 lines

Newsgroups: comp.sys.hp
Path: sparky!uunet!utcsri!helios.physics.utoronto.ca!alchemy.chem.utoronto.ca!system
From: [e-mail] (System Admin (Mike Peterson))
Subject: Re: Problems with 'named'
Message-ID: <[e-mail]>
Organization: University of Toronto Chemistry Department
References: <[e-mail]> <[e-mail]>
Date: Tue, 5 Jan 1993 21:11:34 GMT
Lines: 89

In article <[e-mail]> [e-mail] (Art Harkin) writes:
>  After reading your orignal posting a second time, I still believe that
>a few things you said were unclear, and needed to be corrected. I felt your
>comments implied that include files did not work at all ("if you use $INCLUDE,
>it will destroy your named data, since things like SOA records...")
>I am concerned that this unclear statement could cause other people to
>believe that include files do not work. When in fact they do work, UNLESS
>they put their SOA record for the zone inside the include file (which is not
>a very common thing to do).

I fixed my $INCLUDE files so that the SOA record was in the master file,
and this did make things better, but didn't fix all the problems.
We still got some data corruption. Note that we also use the optional
capability on $INCLUDE to specify the domain name to be appended to all
partially qualified names - my guess would be that that is broken in
the BSD 4.8.3 code too. I was told that the patches I have for named
fixed several unspecified $INCLUDE problems.

>> We have compiled our own named from BSD source, and it works better
>> but not perfectly (the $INCLUDE does work properly however).
>
>  A second thing I believe was misleading was that your comments stated
>this was an HP-UX bug alone, since you said that after compiling BSD sources
>this problem was fixed. This bug is found in BIND 4.8.3 code, and is now
>finally fixed in 4.9 alpha code (not yet officially released by Berkeley).
>You compiled a new named using patches not distributed by BSD.

As I said in my original postings on this, and in the posting you are
quoting from, the identical bugs occur on Domain/OS SR10.4. I therefore
presumed that this was a source code problem, especially since I was
told that the version of named I do run has patches for $INCLUDE.

As for using non-BSD patches to make named work - what else can we do?
We are the master nameserver for our domain, yet none of our HP
systems (Domain/OS or HP-UX) provides a named that works in our
environment. Our users want to use the system, and do things like e-mail
reliably.

>> Given that I originally called in this problem in June, that it was not
>> fixed in 9.0 and that a patch is still not available is none too impressive.
>
>  Your report is the only one filed on this problem for named. After you
>posted on the net a few months ago, I responded to you by private e-mail
>and you mentioned that you already solved the problem by using your own
>named sources and that this part of your named problems was resolved. Thus
>I did not rush a patch for you, since you already fixed it.  A patch is
>forthcoming, but I am delaying it until I can incorporate other patches
>put in BIND 4.9... thus avoiding having a bunch of named patches floating
>around.

Named problems have come up here at least 5 times in the past few
months, and I have sent my named to a couple of people. I have no idea
what exactly was wrong in each case, as named is difficult to set up
from scratch. I know one person definitely had $INCLUDE problems.

HP needs to get on the ball that many/most customers do not call the
ResponseLine even if they are permitted to do so - they don't have time,
don't want to deal with the ResponseLine bureaucracy, don't get (correct)
answers fast enough, whatever.
Many users use comp.sys.hp as their primary line of support
for several reasons: it is very quick, works at all hours/weekends,
accesses a large number of sites with a lot of knowledgable people,
and it is FREE.

On the other hand, I try to document ALL our problems as completely as
possible, and call them all in, so that problems will in fact be
addressed. This has resulted in our site being removed from ResponseLine.

>  Hopefully, you can see that I am concerned that this patch and problem
>be put in perspective. I felt that your original posting was unclear enough
>to let others think this was an HP only problem with include files not
>working at all. This may have cause an uneccessary rush for uneeded patches,
>since they did not understand the problem fully.

I'm not clear what "uneccessary rush for uneeded patches" can happen,
because there aren't any available from HP. If your software is broken,
users with a support contract have the right to expect you to fix it; it
doesn't matter that the problem is / is not caused by someone elses
source code or how difficult the problem is to fix - the customer bought
it from you, with a warranty or support contract that says you will fix
it if it doesn't work.

This problem is critical for us - e-mail/rlogin/ftp don't work when
the nameserver is returning nonsense, and maintaining the data files
without $INCLUDE is very time-consuming and picky work.
--
core error - bus dumped    -*-    Mike Peterson, SysAdmin, U/Toronto Chemistry
 *******   As usual, I speak only for me, myself and I; nobody else   *******
E-mail: [e-mail]  Tel: [phone]  Fax: [phone]