On Thursday 08 January 2009, Josef Wolf wrote:
On Thu, Jan 08, 2009 at 01:17:26AM +0200,
I have /etc/.pwd.lock here dating back to 2006,
so I don't think it's
related to the problem.
When I removed the file, the problem disappeared.
Weird, I still have the file here and groupadd foo returns immediately.
On the other
hand, nscd is very buggy and crashes at least once a day
Hmm, I have no clue why nscd is installed at all... I have not installed
it intentionally. I only learned about its existance while I was tracking
down the problem.
It's meant to be a name service caching daemon, when it doesn't crash that
like a problem with name resolution to me, check
/etc/hosts for the localhost entry.
$ grep 127 /etc/hosts
127.0.0.2 raven.my.domain raven
In order to
investigate this further, try
"strace ls" in an empty directory and check where it spends most time.
Before I wrote my first mail of this thread, I have actually done
strace groupadd foo
and it _was_ actually hanging trying to flock(2) /etc/.pwd.lock. The
problem was perfectly reproducible: I have reinstalled three times
just to figure out where the file came from and to make sure that it
was not me causing the problem.
groupadd foo simply opens /etc/.pwd.lock here, no flock occurs.
Now, I have re-created the file with "touch
/etc/.pwd.lock", and it does
_not_ hang anymore.
Guess, I have to reinstall to reproduce the problem again.
Anyway, a stale lock file should _not_ exist in a clean install. And even
_if_ it exists, this situation should be detected and removed, IMHO.
Thanks for your comment!
If I remove /etc/.pwd.lock, groupadd creates it again. I still don't think the
problem is /etc/.pwd.lock, maybe /etc/ is accessed via NFS?
To unsubscribe, e-mail: opensuse+unsubscribe(a)opensuse.org
For additional commands, e-mail: opensuse+help(a)opensuse.org