As David mentioned, Operations Manager is by far the best option for historical trending, general data gathering and very granular alerting capabilities. There are some Nagios/MRTG/general SNMP plugins out there but they're nowhere near as comprehensive as Ops Mgr. Ops Mgr is also quite reasonably priced -- we encourage almost all customers to get it.
Along with Ops Mgr, there's also Protection Manager which goes towards policy-based replication management, reporting and alerting (much more effectiveway to maintain many SnapMirror or SnapVault relationships than managing the relationships individually).
may I ask what you are using to monitor HOST filesystem utilization? Particularly in block (iSCSI/FCP) environments.
And if you do monitor, do you;
Monitor based on a threshold only? (e.g. Alarm goes off when host filesystem space utilization reaches 80%)
How many hosts do you typically monitor?
Keep a history for all the hosts for trending and reporting?
Do you have a consolidated view of overall filesystem utilization levels?
Thanks in advance,
Phew....that's a big question. Or rather a short question with tons of answers. Most of my experience here has been around Nagios for alerting (with some messing around with ZenOSS and Zabbix) with some Munin for Linux (due its extreme ease of setup and good trending) and some modified Perl scripts for Netware file server trending.
As you can see, it possibly ends up being a different answer as far as setup depending on the OSes involved. (I agree I'd love to see plugins for DFM/Ops Manager for this (expansion of the host agent basically).)
So....some actual thoughts/details would be....
- alerting and trending are two different things as you mentioned and may not be in the same product
- Nagios does scale pretty well -- while not the best example, I had a setup checking 647 hosts (ping checks) with 1447 service checks total running on a 5 year old IBM 1RU server (which was also doing some other things).
- Nagios is incredibly extensible -- my coworker at the time put together a script to pull all the switch IP #'s from his tracking database nightly and monitor them with Nagios (and no, don't get me started on Ciscoworks )
- we did monitor file systems based on a threshold but dependent on what the box was doing (i.e. file servers were bigger so had alerts higher levels, app servers mattered for rapid changes, etc.)
- we did have very good history on the Linux side thanks to Munin but could never come up with something that worked well/consistently across Netware, Linux, Windows and AIX (so had various point solutions or sometimes nothing)
- And no, no consolidated view of filesystem levels....I'd see that as something of a holy grail if you will....but very expensive (whether in time or money....especially if looking for multiple OS support).
So, I'm afraid no magic solutions but hopefully helpful.
I have seen a demo of this http://www.tek-tools.com/ and it looks very good. (They always do...) Have not roled it out on our site however. Cost was not to high either. Have a play for free and see what you think.
Andrew Miller schrieb:
There are some Nagios/MRTG/general SNMP plugins out there but they're nowhere near as comprehensive as Ops Mgr.
May I add, that I have developed several Nagios plugins which are based on ZAPI-calls and allow a more comprehensive monitoring and trending. F.e. latency per volume, usage of all volumes witch one check (dynamic identification of newly created or deleted volumes, …). While not free, they are actively developed and well tested. More information on the website http://www.netapp-monitoring.info/en/