Forum: CFEngine Help
Subject: Re: 180+ instances of cf-execd and cf-agent
Author: davidlee
Link to topic: https://cfengine.com/forum/read.php?3,26261,26265#msg-26265

What version of cfengine?  This might matter... see below.

Over the last 15 months or so, our cfengine3 installation has grown from zero 
to about 30 machines (RHEL 5.x), running version 3.1.4 of cfengine-community.   
I have seen such a pile up of cf-execd and cf-agent processes on one occasion, 
and it was on a heavily loaded machine. So for us it was a rare event.

I suspect that one of the internal locking databases (BerkeleyDB) had got 
corrupted.  Simply restarting cfengine (with a clearout of residual processes) 
didn't help; the processes started to pile up again.

I can't remember the exact cure, but I think it involved wiping the various 
"/var/cfengine/state/*.db" files between stopping and restarting cfengine.

cfengine has moved on since 3.1.4; it is now at 3.3.4.  One of the major 
changes has been to use a different database backend, not BerkeleyDB.  Indeed, 
I understand that a major reason for this decision was precisely BerkeleyDB's 
poor resilience under certain load and error conditions.  For problems such as 
the one I saw (which sounds like yours), the advice would almost certainly be 
to migrate to a newer version of cfengine, rather than spend too much time 
trying to debug this problem on this older version with its known-problematic 
backend database.

(Your mileage may vary, of course!)

_______________________________________________
Help-cfengine mailing list
Help-cfengine@cfengine.org
https://cfengine.org/mailman/listinfo/help-cfengine

Reply via email to