Server_stats from celerra side or symstat/performance manager from DMX

Thanks. Would running a server_stat during peak times bear any impact ?

I've never encounter any issues running it during peak times

No, the server_stats command set will not impact performance of the system.

It looks like server_stats allows real time monitoring.  How can I collect stats over a time frame?  I have an immediate need to discover IOPS rates for a particular file system in order to determine if it can be moved to SATA drives.

I need to look at a history over a week, or so.

Thanks...

Try celerra monitor

Sent from my Verizon Wireless 4GLTE Phone

We have a recurring issue with  /nas/jserver filling up. 

The result is that Celerra Monitor will not display stats.

EMC Support has not been able to suggest a fix for this.  This file system cannot be extended.

It appears to be a design flaw in Celerra that EMC will not correct.

Any other ideas?

Thanks!

Thank you. 

It looks like server_stats allows real time monitoring.  How can I collect stats over a time frame with server_stas?

Server_stats or Navi/UniSphere Analyzer

You can use server_stats with –file –format csv to write the values into a CSV file and then use Excel to analyze it – see man server_stats

Rainer

Thanks for your info

Yes details would be nice

We run a number of cron jobs on the control station to collect long term stats. We write some to disk and send some to an external stats collection service called graphite ( http://graphite.wikidot.com/ ).

This command will run every 5 minutes and collect diskVol performance stats:

*/5 * * * * /nas/bin/server_stats server_2 -monitor diskVolumes-std -i 5 -c 60 -file "/nas/quota/slot_2/vg8_stats/diskVol/server_2/server_2_`date '+\%FT\%T'|sed s/://g`" > /dev/null

We use this to grab top_talkers every 5 minutes as well:

*/5 * * * * /nas/bin/server_stats server_2 -monitor nfs.client -i 5 -c 60 -file "/nas/quota/slot_2/vg8_stats/top_talkers/server_2/server_2_`date '+\%FT\%T'|sed s/://g`" > /dev/null

Here's a sample of the presentation of our key latency stats that we export to graphite:

If folks are interested in our work in intergrating server_stats with graphite, let me know.

If you could send on details it would be great.

Jeff,

Very nice.  Yes, one more vote for a write-up on how to integrate it as you have.  Looking forward to it if you have time.  Thanks!

hi Jeff could you guide me please how to implement how to transfer perf stats/external stats collection service called graphite ( http://graphite.wikidot.com/ )