I've created a Zenoss Core 3.1 machine as an EC2 instance, and am trying to get the Linux, Linux Addon, and Ubuntu zenpacks working for monitoring. The issue I'm having is with zencommand timing out when trying to gather data. Specifically, it times out on the uptime, /proc/stat, and /proc/meminfo commands (although I don't really see any data whatsoever in the reports). This happens both when trying to monitor the Zenoss machine itself as well as a separate dev EC2 instance.
Things I've done:
The output I'm seeing:
Any help would be appreciated. Zenoss seems to be the exact monitoring system we were looking for, but this issue has me stumped.
This is interesting - in the Zenoss web interface, if I go to Advanced->Monitoring Templates, then pick the /Server/SSH/Linux class to see its data sources, I can "test" the cpu, mem, and uptime commands. In this interface they all work perfectly. So there's a difference in the way they're being called from the web UI vs. the actual daemon. Does that sound familar to anyone?
Sent from my HTC Touch Pro2 on the Now Network from Sprint®.
It looks like I figured out the problem. In case others run into this, the issue was that I hadn't set the zKeyPath to point to the proper EC2 keyfile (it was still on the default of id_rsa). I had set the IdentityFile in /home/zenoss/.ssh/config properly, which allowed the Test of the reporting commands to run successfully (as I posted, ssh-ing worked fine from the commandline with no keyfile prompt), but I guess zencommand uses a different set of parameters. It would be nice if the Test and the actual command used the same execution parameters, but whatever, it works now.
To inform everyone a bit here - you always need to test by running the particular daemon in verbose mode - maybe in the foreground or just to the log temporarily to see what's really going on. It is unfortunate, and a bit of a bug I think, that the various other ways to run commands do not replicate the environment or necessarily options used by the daemons. Of course, this has been like this since 1.1, so I have my doubts it's going to be addressed soon.
Information Technology Area Supervisor
LEPP Computer Group
Follow Us On Twitter »
||Latest from the Zenoss Blog »||Community||Products||Services||Customers||About Us|
Copyright © 2005-2011 Zenoss, Inc.