I would like to get performance data from weekly PERFORMANCE DATA autosupport message. I found files with ccma extension in the asup but I don't know how to open it. Is it even possible for a user to get something from this asup messages?
thank you for reply. I don't see CM-STATS-HOURLY-DATA.xml file in the asup that I receive in email. I checked Active IQ and the file is not there too. Only CM-STATS-HOURLY-DATA-01.TAR if I open the PERFORMANCE DATA asup.
Under ONTAP, there is a sub-system called 'Counter Manager archiver process' which generates this file, depending upon the load on your cDOT node, it can delay the file and ASUP process may time-out for this file. Once the Counter Manager archiver process is less busy, the hourly statistics data file will be correctly generated.
.tar (Performance stats) : is parsed once per day by the AutoSupport subsystem and reported in the XML format. If this parsing issue occurs early in the process, then no XML data is returned by the AutoSupport subsystem.
I need create monthly report with performance data like latency, IOPS, CPU utilization. In the past I used Active IQ graphs but these are just terrible. With gaps and values out of the bounds (mistakes). I reported this several times during the last 2 or 3 years but it was never fixed and the cases just ended up in void.
The best for me would be to get raw data. Direcly from storage, from asup messages, Active IQ or any other way. We also use NAbox (appliance with harvest-graphite-grafana) but it is not supported tool that can end or break any day so I cannot rely on it.
So the basic question is how to get the raw performance data?
1) First of all check : How your cDOT node's are doing interms of processing the autpsupport files:
Enter the command as I showed below: Following exmaple, shows 'completed' that means no errors. What's your output ?
This is for 1st Node, but you can check all the Nodes.
SANCL1::*> system node autosupport manifest show -node SANCL1-01 -body-file cm_* AutoSupport Collected Node Sequence Body Filename Size Status Error ----------------- --------- -------------- ----------- --------- ----------- SANCL1-01 24xx cm_stats_hourly_info.xml 331.5KB completed cm_stats_hourly_data.xml 305.5MB complete
It all depends on the 'load' on the system nodes, sometimes these files simply times out or space out.
2) Are you using OCUM/AUM : What version is that. I use OCUM for IOPS/MBPS health monitoring and it works great for me. What is the issue you are facing ? Which setting or graph you are looking for ?
Active IQ Unified Manager : 9.6P1 is a great too, however it will not straight away give you month's data. However, I am interested to know what version of OCUM you are running. I can see monthly data statistics on my OCUM. Let us know exactly what is terrible.
3) Raise a ticket with NetApp for that BUG: Most likely they will check the overall health of your system and then decide if you are really hitting the bug.
4) As final option : We have all commands, thanks to QoS statistics, we can access the performance of the Cluster, Node, Volume, LUN, LIF etc.
Active IQ Unified Manager can definitely do all of what you need. If you have an old OnCommand Unified Manager instance, it can be upgraded as well. The main thing is that the performance data will be imported to a few days prior, but won't go back historically months unless you had an existing OCUM install.
If you need help doing so, feel free to let us know.