Can anyone tell me just how robust zencommand really is?
Will it EVER kill its dead children? (sounds awful to ask it like
that)
How many checks are reasonable?
Should I put a usleep delay in my commands to slow them down for
Zenoss?
currently I have changed the conf file to commandTimeout 6, and
changed the zencommand.py to a max of 512 connections with no relief.
I've seen zencommand running around 1500 commands on a 5 minute
schedule without the problems you've described. It doesn't seem right
that something as simple as a name resolution would cause a problem.
If you see the data returned very quickly to the plugin, but it
doesn't write it and exit immediately I would look at your disk to
make sure you don't have a bottleneck there.
If you look around line 118 in $ZENHOME/Products/ZenRRD/zencommand.py
you will see that it does send the KILL signal to processes that have
exceeded their timeout.
Sleeping in your scripts wouldn't help alleviate the problem. It
would likely only make it worse.
As far as the command timeout goes, you should set
zCommandCommandTimeout in the zProperties. I believe the default is
15 seconds.
_______________________________________________
zenoss-users mailing list
[email protected]
http://lists.zenoss.org/mailman/listinfo/zenoss-users
