Getting current NFS OPS / Cache Hits via SMTP?

[ALEJANDROB]

Hi1 i'm trying to install some nagios alerts to be triggered when NFS OPS are high or Cache Hit goes down. I know that i can get some information via smtp.

NFS OPS  -  SNMPv2-SMI::enterprises.789.1.2.2.27.0

Cache Hits  - SNMPv2-SMI::enterprises.789.1.26.8.0

Cache Misses -  SNMPv2-SMI::enterprises.789.1.26.9.0

But those stats are since filer was booted. I need something more real time.

Any clue about this?

Thanks.

[mharding]

Hello - did you ever get an answer to your question?  For NetApp system Alerting have you tried Unified Manager?  The former Operations Manager capability provides automated alerting, and it's free with your hardware purchase.  Along with that comes Performance Advisor, which enables very granular system Performance-based thresholding and alerting.  Also free.  You can scan the Admin guides on the Support download page to see if it does what you want, and the download for that single package -- OnCommand Unified Manager Core v5.1 --  is also available through the Support site.  http://support.netapp.com/documentation/docweb/index.html?productID=61474

[JORDAN_SLINGERLAND]

Your first problem is that you are confusing SNMP and SMTP.

Your second problem is that in your NMS you are using a guage variable and not a counter.

It does not matter if the value is ops since last boot if you take the difference between the current value and the last value you received.

--Jordan

[bobshouseofcards]

As Michael indicates OnCommand Unified Manager certainly will meet the need for the Filer alert and most every other Filer reporting/alerting need you may have.  If you are using Nagios to monitor a bunch of discrete hardware/systems and you just want to add in this item, there is not an SNMP counter in the NetApp MIB that directly provides what you want.  All the NFS statistics are "since last reset", so you have two basic options to make this work.

1.  Get counter values over time and derive the rate in your collection tool based on the differences.

2.  Get counter values and immediately reset the statistics via "nfsstat -z" as a command to the Filer.  Due to the natural lag between query and reset there will be some loss of signal in this mechanism of course.

I've found that most all collected statistics available are every increasing counters as they are efficient to implement at the Filer, leaving any rate/per type calculations to the management tool's resources.

[JORDAN_SLINGERLAND]

Yes. You could also configure your thresholds in On Command and have that send an SNMP trap to your NMS.  This way you get the best of both worlds.  You do not need to reinvent the wheel and can use all the built in thresholds,  yet you can view your aggregated alerts from a single pane of glass.