I agree, I would have prefered that bug 148982 was "fixed" by changing sysstat to show AVG rather than changing SNMP to show ANY.

Since sysstat has the -m option which allows viewing AVG as well as individual CPU performance, it seems reasonable that the same data should be made available via SNMP.

I too have not yet upgraded beyond 6.4.5, and am happy to have seen this discussion prior to the upgrade so I won't be surprised and waste a lot of time trying to understand where the CPU usage increase came from.

Re:

From: Davin Milun milun@cse.Buffalo.EDU Date: Wed, 15 Mar 2006 08:23:54 -0500 To: toasters@mathworks.com Subject: Re: CPU Stats on multi-proc filers

At 7:57, on Mar 15, 2006, Davin Milun wrote:

...

...

We noticed a 20% increase in CPU usage when we upgraded from 6.4.4P6 to 6.5.4P1. First we were worried because there was little chance that this was due to changed usage patterns. Then we investiged by comparing sysstat output to snmpd output. Our conclusion was that in OnTAP 6.4.4P6, snmpd reports cpu utilization as the average of each CPU's utilization. With the upgrade to 6.5.4P1 snmpd reports CPU utilization as percentage of time when at least one CPU is busy, i.e. ANY instead of AVG. That has not changed when we upgraded to OnTAP 6.5.6P3.

Oh.

I'm still running 6.4.x on the filers that I watch CPU most closely on.

So if NetApp broke cpuBusyTime in 6.5.x to report ANY time, then (a) I stand corrected, and (b) I'm quite unhappy with that change.

And I just did some manual testing, and it's indeed the case that BugID 148982 caused NetApp to break the SNMP values in 6.5.4! :-(((

So the once-useful SNMP cpuBusyTime is now showing the ANY time.

They should have fixed that bug by fixing sysstat to show AVG, not by fixing SNMP to show ANY!

Thanks for the heads up! I would have had the same "What's causing this 20% CPU spike!?!?" issue after my upcoming upgrades, if I had not had this discussion here.

Davin.