[OpenAFS] fetchmail / afs / segfault revisited

[John Bleichert]

Fetchmail seems to work fine, now (2.4.17-xfs) - my problems seem to be 
in/with procmail which I'm running to split up all my mailing lists into 
folders in Pine.

Apparently there are internal setuid/gid issues to contend with. I'm 
working on it (or at least *will* be when I get home from work!).

JB

 On Mon, 18 Mar 2002, Pavel Semerad wrote:

> Date: Mon, 18 Mar 2002 10:31:03 +0100
> From: Pavel Semerad <semerad@ss1000.ms.mff.cuni.cz>
> To: Johnny B . <syborg@stny.rr.com>
> Cc: openafs-info@openafs.org
> Subject: Re: [OpenAFS] fetchmail / afs / segfault revisited
> 
> 
>   There were a problem with linux kernels >= 2.4.10 . When file
> was open with O_SYNC flag (as fetchmail does), write to this file
> caused a segfault. I only know that current CVS version is OK
> (tested with kernels 2.4.17, 2.4.18). I don't know who and when
> did repair it.
> 
> Pavel Semerad
> 
> > Hello All
> > 
> > I'm trying to run fetchmail from my AFS account and it segfaults
> > immediately. I mentioned this before but couldn't find anyone having
> > similar failures :-(
> > 
> > This time, I may have found the fault using strace, but I'm not sure
> > how to interpret it. It looks like fetchmail may not be running with
> > my token, even though it's running from my ~/bin. Here is the end of
> > the strace output:
> > 
> > <...>
> > umask(077)                              = 02
> > rt_sigaction(SIGABRT, {0x8053a40, [ABRT], SA_RESTART|0x4000000},
> >    {SIG_DFL}, 8) = 0
> > rt_sigaction(SIGINT, {0x8053a40, [INT], SA_RESTART|0x4000000},
> >    {SIG_DFL}, 8) = 0
> > rt_sigaction(SIGTERM, {0x8053a40, [TERM], SA_RESTART|0x4000000},
> >    {SIG_DFL}, 8) = 0
> > rt_sigaction(SIGALRM, {0x8053a40, [ALRM], SA_RESTART|0x4000000},
> >    {SIG_DFL}, 8) = 0
> > rt_sigaction(SIGPIPE, {0x8053a40, [PIPE], SA_RESTART|0x4000000},
> >    {SIG_DFL}, 8) = 0
> > rt_sigaction(SIGQUIT, {0x8053a40, [QUIT], SA_RESTART|0x4000000},
> >    {SIG_DFL}, 8) = 0
> > open("/afs/vonbek.dhs.org/usr/johnnyb/.fetchmail.pid",
> >       O_WRONLY|O_CREAT|O_EXCL|O_SYNC, 0666) = 3
> > getpid()                                = 7695
> > write(3, "7695", 4 <unfinished ...>
> > +++ killed by SIGSEGV +++
> > 
> > 
> > I've cleaned it up a little to deal with the hard wrap I use in mutt.
> > It doesn't seem like write() is failing, it seems like write() doesn't
> > even get to run. It looks like it's being interrupted, correct ?
> > 
> > Any insight would be appreciated!
> > 
> > JB
> > 
> > +---------------------------------
> > | John Bleichert                       
> > | syborg@stny.rr.com
> > | http://vonbek.dhs.org/latest.jpg
> > 
> > _______________________________________________
> > OpenAFS-info mailing list
> > OpenAFS-info@openafs.org
> > https://lists.openafs.org/mailman/listinfo/openafs-info
> _______________________________________________
> OpenAFS-info mailing list
> OpenAFS-info@openafs.org
> https://lists.openafs.org/mailman/listinfo/openafs-info
> 

-- 
|-John Bleichert----syborg@stny.rr.com----------------|
|-------------------http://vonbek.dhs.org/latest.jpg--|