[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Hung CVS process

From: Larry Jones
Subject: Re: Hung CVS process
Date: Thu, 18 Sep 2003 15:17:14 -0400 (EDT)

Rick Genter writes [using very long lines]:
> I have an OpenBSD 3.3 system that I am using as a CVS server, with
> several clients connecting using WinCVS from Windows 2000 and Windows XP
> machines. Today one of my users did "something" that ended up with a
> hung CVS process on the server. By hung I mean that when I looked it had
> accumulated over 2 hours of CPU time, was using 99% of the CPU, and the
> CPU was spending 14% in User space and 85% in System space. Has anyone
> seen anything like this before? Any idea of a cause? 

This happens when CVS crashes in the cleanup routine -- the signal
handler calls the cleanup routine which just crashes again ad infinitum
(or at least until you run out of stack space).

> I ended up having to kill -9 the process (no response to kill -TERM or
> kill -HUP). I'm concerned that the repository may be in an inconsistent
> state, or that there may be lock files hanging around. How do I check
> the integrity of the repository?

There shouldn't be any inconsistency in the repository, although you can
use the contrib/check_cvs script to check.  There probably are stale
lock files hanging around though -- something like:

        find $CVSROOT -name '#cvs*' -print

will find all the locks (unless you're using LockDir= in your
CVSROOT/config file, in which case use that directory instead of
$CVSROOT), then you can remove any that are stale.

-Larry Jones

I think grown-ups just ACT like they know what they're doing. -- Calvin

reply via email to

[Prev in Thread] Current Thread [Next in Thread]