[Linux-HA] Debugging CRM Resource Exit Codes?
alanr at unix.sh
Mon May 15 14:42:52 MDT 2006
Brian A. Seklecki wrote:
> On Sun, 14 May 2006, Alan Robertson wrote:
>> Andrew Beekhof wrote:
>>> The required exit codes are listed at:
>>> * http://linux-ha.org/HeartbeatResourceAgent
> My biggest concern is that it's not an entirely uncomplicated task to
> debug the decision making process by reviewing the log files.
> You have consider that this entire code base is essentially a very lage
> decision making aparatus designed to exec shell scripts to move
> resources around machines, right?
> Verbosity is almost always a good thing, but if you can't very easily
> ascertain why a resource is in a current state, then the purpose of the
> software is mitigated.
> As an admin (user I suppose), I need to be able to easily run a command
> like sbin/crm_mon or sbin/crm_resource and easily see that
> $prefix/etc/ha.d/resourced/IPaddr failed *because* it was running as a
> user with insufficient priviledges to exec or because it couldn't fork
> or because the ...
These typically would be logged by the resource agent itself. Anything
the resource agent writes to stdout or stderr we also log (through the
> Right now, running heartbeat(8) with -d (which presumably passes "-d" to
> all child processes it exec's), I can't even see that.
I don't know about your logging setup. You need to read your syslog
configuration. You might also look at /var/log/localmessages. This
kind of thing is different from one version of Linux to another.
Probably we should consider changing our default from local7 to daemon
or something. It probably got set to local7 because that's what we
typically run with while testing - because it helps us our test
automation software to log to a resource no one else is logging to.
Alan Robertson <alanr at unix.sh>
"Openness is the foundation and preservative of friendship... Let me
claim from you at all times your undisguised opinions." - William
More information about the Linux-HA