On Sun, Jun 18, 2006 at 03:10:56PM +0200, Bernd Schubert wrote:
> >> Just try 'lsof |grep mothermole'.
> >
> > This hangs, just as my try with fuser did.
>
> Hmm, this sounds as if /mothermole is in your PATH, LD_LIBRARY_PATH or
> whatever. Can you check this? If you don't find anything, try to strace
> lsof and see where it hangs.
No, it's definitely neither in my $PATH nor in my $LD_LIBRARY_PATH.
The strace ends with:
rt_sigaction(SIGALRM, {SIG_DFL}, {0x80573f8, [ALRM], SA_RESTART}, 8) = 0
rt_sigaction(SIGALRM, {0x80573f8, [ALRM], SA_RESTART}, {SIG_DFL}, 8) = 0
alarm(15) = 0
write(5, "Cr\5\10", 4) = 4
write(5, "\f\0\0\0", 4) = 4
write(5, "/mothermole\0", 12) = 12
write(5, "\0\20\0\0", 4) = 4
read(6, "\377\377\377\377", 4) = 4
read(6, "\26\0\0\0", 4) = 4
read(6, "\f\0\0\0\0\0\0\0\0\0\0\0\250\26\0\0\260a\0\0\1\0\0\0\0"..., 4096) = 4096
alarm(0) = 15
rt_sigaction(SIGALRM, {SIG_DFL}, {0x80573f8, [ALRM], SA_RESTART}, 8) = 0
rt_sigaction(SIGALRM, {0x80573f8, [ALRM], SA_RESTART}, {SIG_DFL}, 8) = 0
alarm(15) = 0
write(5, "dr\5\10", 4) = 4
write(5, "\f\0\0\0", 4) = 4
write(5, "/mothermole\0", 12) = 12
write(5, "`\0\0\0", 4) = 4
read(6, 0xafea7b98, 4) = ? ERESTARTSYS (To be restarted)
--- SIGALRM (Alarm clock) @ 0 (0) ---
alarm(0) = 0
rt_sigaction(SIGALRM, {SIG_DFL}, {0x80573f8, [ALRM], SA_RESTART}, 8) = 0
close(5) = 0
close(6) = 0
rt_sigaction(SIGALRM, {0x80573f8, [ALRM], SA_RESTART}, {SIG_DFL}, 8) = 0
alarm(5) = 0
wait4(-1,
So there are some system calls with /mothermole as argument, and then it
hangs, as expected. Can I still somehow find out, which processes are
waiting?
> > That b0rked up NFS mount seems to cause a lot of trouble, i.e. updatedb
> > processes from the last view days are hanging around, blocked, and also
> > some other programs refuse to start as the try to access that
> > mountpoint.
> >
> > Is there really no other way then rebooting to get rid of this nasty
> > thing?
>
> Processes in D state are difficult to kill. If nothing helps, you could put
> any other system on the nfs server ip address and export a directory
> as /mothermole, the processes should die with a stale nfs filehandle.
I guess that would do the trick. I could also bring up the NFS server
again and export a dummy share (the real share behind /mothermole was a
RAID-5 array that is currently broken btw.) But I really only want to do
this as a last resort.
It is really annoying that a NFS client can be brought into such
troubles and there is no sensible way to resolve it.
Greetings,
Frank
--
Frank Blendinger | fb(at)intoxicatedmind.net | GPG: 0x0BF2FE7A
Fingerprint: BB64 F2B8 DFD8 BF90 0F2E 892B 72CF 7A41 0BF2 FE7A
Attachment:
signature.asc
Description: Digital signature