[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Stale lock files
From: |
Dan Peterson |
Subject: |
Stale lock files |
Date: |
Mon, 17 Jun 2002 13:27:58 -0700 (Pacific Daylight Time) |
We often find ourselves with stale lock files hanging around and causing
problems. (pserver, SunOS-5.8, cvs-1.11 and just recently upgraded to
1.11.2).
After poking around in the code a bit, I noticed the #define for
"CVS_FUDGELOCKS", which will remove locks older than CVSLCKAGE seconds.
I was thinking about turning this on... probably with a CVSLCKAGE longer
then the default of 1 hour (maybe more like 4-8 hours?) and was wondering
what the adverse consequences might be?
The only real problem I can see is if the process which created the lock
is still running.
Then I thought maybe it wouldn't be too tough to perform an extra check to
see if the process is still running before removing the lock. This would
be fairly easy (assuming /proc exists) for the read/write locks but much
tougher for the master lock. The only way I can see to do this for the
master lock would be to first check for the existence of a read/write lock
in the same directory then see if the file ctime is similar (within a few
seconds?) of the master lock. If so, assume the master was created by the
same process that created the read/write lock.
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- Stale lock files,
Dan Peterson <=