Options
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Mute
- Printer Friendly Page
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I am working on a plan to upgrade from OnCommand Unified Manager 5.1 to 5.2R1. On page 11 of the Release Notes, it says:
------------------------------------------
CPU monitoring is disabled by default on a newly installed DataFabric
Manager server. The CPU Usage (%) graph data is not available.
CPU monitoring is also disabled by default if you have upgraded to
OnCommand Unified Manager Core Package 5.2R1. The data in the CPU
Usage (%) graph expires over a period of time.
The historical data is available in the CPU Usage (%) graph in both the
OnCommand console and Operations Manager console until they reach
expiry.
------------------------------------------
Why would this be the case? I am tracking a CPU issue on one of our filers now and need this information! If I am reading correctly, I will not be able to view Processor information in Performance Manager after the upgrade. Is this accurate?
Solved! See The Solution
1 ACCEPTED SOLUTION
migration has accepted the solution
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Greetings,
The public report for the bug where this change occurred states the option:
http://support.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=612203
CPU usage can be tracked through performance counters collected by Performance Advisor. Use the CPU Utilization graph in "Storage System Summary View" to track the CPU utilization trend.
From 5.2 the CPU monitor is disabled by default, the monitor can be enabled by running following command:
# dfm option set cpumoninterval=
You should also find this information in the Release Notes starting with UM 5.2GA.
Thanks,
Kevin
8 REPLIES 8
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
Nothing is changed, only CPU monitoring is turned off by default. This was done intentionally because some of the customers wanted to run the filers at high CPU but were bugged with the warning events.
You just need to turn on the CPU mon by changing the option to Yes after upgrade for CPU monitoring to start working.
Thanks,
Arun
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks Arun! So I understand that CPU monitoring is now turned off by default, but I just want to verify that I can still use Performance Monitor to view data about the CPU. The line "The CPU Usage (%) graph data is not available" is what concerned me. That makes it sound like you can't chart CPU data anymore. Is that accurate?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
AFAIK, once you enable to the CPU mon the CPU graph should start showing up.
Thanks,
Arun
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Thanks Arun! I have looked through the Ops Manager and Performance Advisor documents, and I cannot find how to manually enable CPU graph data. Do you know what the command or menu option is?
migration has accepted the solution
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Greetings,
The public report for the bug where this change occurred states the option:
http://support.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=612203
CPU usage can be tracked through performance counters collected by Performance Advisor. Use the CPU Utilization graph in "Storage System Summary View" to track the CPU utilization trend.
From 5.2 the CPU monitor is disabled by default, the monitor can be enabled by running following command:
# dfm option set cpumoninterval=
You should also find this information in the Release Notes starting with UM 5.2GA.
Thanks,
Kevin
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Very helpful, thank you kryan!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
This is true...
Since the newer versions of OnTap (8.1+) changed the way multiple CPUs cores are utilized, its somewhat broken the normal methods of seeing CPU utilization like DFM and sysstat.
Our 6210 with 8 cores per head can show being pegged at near 100% cpu in DFM for most the day, however in reality its between 30-40% - we get a few annoying "cpu too busy" messages a day.
To get a more accurate reading of total CPU usage you have to use the "sysstat -m" command from the controller's CLI and look at the Avg column:
Filer> sysstat -m 1
ANY AVG CPU0 CPU1 CPU2 CPU3 CPU4 CPU5 CPU6 CPU7
97% 43% 45% 41% 40% 41% 44% 45% 44% 44%
97% 41% 44% 38% 38% 40% 42% 41% 43% 42%
96% 40% 42% 37% 38% 40% 42% 42% 42% 41%
100% 59% 57% 56% 55% 56% 61% 62% 60% 61%
99% 48% 50% 47% 44% 47% 51% 50% 50% 49%
99% 55% 55% 52% 53% 51% 57% 56% 57% 56%
96% 37% 38% 35% 36% 36% 39% 37% 39% 38%
"Any" is the metric DFM uses to show CPU usage - and as you can see above, its not really representative of actual filer load.
Furthermore - background processes will consume cpu, but will back-off in response to client workloads. so without looking at other metrics like throughput and latency, CPU% is rather irrelevant.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Very helpful to know! Thanks Colin.
