Strange signals and cfengine fails to run.

[Russell Adams]

I'm having issues with cfengine failing to run, coinciding with syslog
messages on the affected host.

It doesn't look like cfagent is doing anything, including updates
while this is occuring (no email or outputs).

I'm running cfengine 2.0.8_p1 on Gentoo, with cfexecd as a direct
respawning child of init in foreground mode. That way if cfexecd were
to die, it would be restarted. Cfengine is responsible for making sure
other processes are running, so needs a direct link back to init to
ensure it runs.

>From inittab:
# CFEngine
cf:5:respawn:/var/cfengine/bin/cfexecd -g

Here's the syslog messages:

Nov 10 10:02:43 soja grsec: signal 11 sent to (cfagent:29504) UID(0)
EUID(0), parent (cfagent:14008) UID(0) EUID(0)
Nov 10 10:02:43 soja grsec: signal 11 sent to (cfagent:14008) UID(0)
EUID(0), parent (cfexecd:25212) UID(0) EUID(0)

It sounds like cfexecd is killing cfagent, but I'm unsure as to why.

I've seen this behavior on two of my systems now.

The only thing I can think of is that perhaps cfexecd isn't reacting
properly when cfagent.conf and update.conf are updated from the remote
server. Could it be hanging onto an old file pointer and trying to
open the deleted file?

The behavior persists until cfexecd is killed and a new one started.

Any input is appreciated.

Russell