Am I going to run out of locks soon?

Gary Mills mills at cc.umanitoba.ca
Fri Oct 7 09:10:37 EDT 2005


On Thu, Oct 06, 2005 at 05:57:25PM -0400, Igor Brezac wrote:
> 
> On Thu, 6 Oct 2005, Gary Mills wrote:
> 
> >On Thu, Oct 06, 2005 at 05:00:21PM -0400, Igor Brezac wrote:
> >>
> >>On Thu, 6 Oct 2005, Gary Mills wrote:
> >>
> >>>Take a look at the first and third numbers here.  The first number
> >>>has been increasing.  Will something break when it hits 50000?
> >>>Or, are those numbers not related?
> >>>
> >>>	# su cyrus -c '/usr/local/BerkeleyDB.3.1/bin/db_stat -c'
> >>>	47690   Last allocated locker ID.
> >>>	3       Number of lock modes.
> >>>	50000   Maximum number of locks possible.
> >>>	0       Current lockers.
> >>>	0       Maximum current lockers.
> >
> >>This used to be an old bug in cyrus which was fixed quite awhile ago (at
> >>least it was fixed for db4).  What version of cyrus-imapd do you use?
> >
> >Hmm, interesting.  We're still running cyrus-imapd-2.1.14.
> >I'm planning on upgrading, but not this month.
> 
> Hmm, this was fixed in 2.1.14: 
> https://bugzilla.andrew.cmu.edu/cgi-bin/cvsweb.cgi/src/cyrus/master/service.c.diff?r1=1.42&r2=1.43

This turned out to be a false alarm:

	# su cyrus -c '/usr/local/BerkeleyDB.3.1/bin/db_stat -c'
	51608   Last allocated locker ID.
	3       Number of lock modes.
	50000   Maximum number of locks possible.
	0       Current lockers.
	0       Maximum current lockers.

Diligent reading of DB source revealed that the two numbers are
not related.  The first is simply a unique number that's given to
each locker.  The second is the size of a table.

-- 
-Gary Mills-    -Unix Support-    -U of M Academic Computing and Networking-



More information about the Info-cyrus mailing list