We have a two jump boxes that are externally facing...we just use them for ssh access. I have both servers individually added to zenoss. I want to setup some advanced alerting so that an email goes out if, and only if, zenoss cannot make an ssh connection to BOTH of them.
I was going to set this up similar to a polyserve nfs cluster we have setup, but it won't work properly. With the polyserve cluster, they have floating ip's. Zenoss binds the the ip address of the device at the time it's added. Subsequent remodels does not change the ip address on the main device page. So the pinged ip address for the device remains static unless manually changed by the user. For a cluster with a floating ip address...this is fine because the ip will always float to one of the available servers. As long as one of the servers is available....I can configure zenoss so that it won't falsely alert that the cluster is down. With ssh in this case...I can't really do this. If our round robin dns entry resolves to server A and it goes down....zenoss will think the cluster is down when server B may be readily available to server ssh sessions. I'm trying to think of ways to only alert when BOTH of these ssh servers are unavailable. So far, the only thing I can think of is doing some custom scripting in a script on the zenoss box and setting up a command template. The custom script would check to see if at least one of them is available...and if not return a critical alert. Or perhaps setup a custom transform. However, both of these methods seem like they could be very time consuming and error prone. Is there an easy way to setup some kind of even correlation to only send an email if all nodes are down for service? -------------------- m2f -------------------- Read this topic online here: http://forums.zenoss.com/viewtopic.php?p=31598#31598 -------------------- m2f -------------------- _______________________________________________ zenoss-users mailing list [email protected] http://lists.zenoss.org/mailman/listinfo/zenoss-users
