Being a relatively new user to APM/SAM, I was surprised to find that there is no pre-build method for handling mount points. So I did some poking around, and stumbled across http://thwack.solarwinds.com/docs/DOC-71573 by Ohmus. It seemed to cover what I needed, with one issue, I'd have to create a new custom monitor for every mount point. A SQL cluster in our environment has at least 3 mount points for SQL (log, DB, and tempdb), or more depending on the DB requirements and performance needs. Then comes our DFS servers, which have many more mount points (one has 26 mount points for example). This clearly becomes unmanageable, especially when we have upwards of 30 SQL clusters, and 5 DFS servers. So I threw some crude PowerShell together to pull all the volumes via WMI (see attached code). This basically calls WMI looking for all volumes with no drive letter. This works great, except when we hit the servers with more than 10 mount points. SAM's forced limit of 10 script outputs means we cannot "see" what the volume data is, but we can still get alerted on it (script exit code). This makes it impossible to observe any trending.
So 3 main questions;
- Is there any way to increase the 10 output limit?
- If not, what other methods can we use to monitor volumes, and large volume counts?
- Because the mount point names are used for the Message. Statistic. output values (because they are different from role and host perspectives), what does SAM do when they are different from node to node?