[2225] in Release_7.7_team
Re: named on 8.4 Irix randomly dies?
daemon@ATHENA.MIT.EDU (Karl Ramm)
Sat Apr 22 14:51:07 2000
To: Greg Hudson <ghudson@MIT.EDU>
Cc: Garry Zacheiss <zacheiss@MIT.EDU>, release-team@MIT.EDU
From: Karl Ramm <kcr@1ts.org>
Date: 22 Apr 2000 14:50:58 -0400
In-Reply-To: Greg Hudson's message of "Fri, 21 Apr 2000 18:22:03 -0400"
Message-Id: <uusr9by0y25.fsf@ORTHANC.MIT.EDU>
Greg Hudson <ghudson@MIT.EDU> writes:
> > Twice now, I've discovered that named on my 8.4 Indy
> > (science-fiction-double-feature, an R4600XZ) has died for no
> > apparent reason. It doesn't appear to be logging anything
> > interesting or leaving a core dump anyplace.
>
> This is, of course, the same named code base that we backed out of the
> 8.3 release in favor of BIND 8.1. "Oops." (Although I don't think we
> saw this particular failure mode.) I'll do the backout on the
> mainline so we don't have the same problems in 8.4 as we did in 8.3.
Someone ought to quickly test 8.2.2p5 to see if it has the same problem.
(sadly, it would have to be someone who actually uses hesiod (I don't
anymore), or I'd have a reading for you (it's what I'm running at home))
Given the problems we saw, it would be sufficient for someone to just build
it and run it on a regularly used machine.
Hmm, what version of bind are the linux machines running?
(in the long term, someone ought to test bind 9 similarly, perhaps once
it's out of beta.)
kcr
