Unsolved
This post is more than 5 years old
5 Posts
0
452
March 13th, 2012 04:00
Celerra NS20 slowness
Hi,
I have a NS20 with only one Cifs server with 2.2Tb for photo images. I have two machines serving this photos and sometimes the response of the celerra is slow.
With server_stats I see this when the slowness happen:
| server_2 | CPU | Network | Network | dVol | dVol |
| Summary | Util | In | Out | Read | Write |
| 12:45:20 | 93 | 403 | 1084 | 45664 | 432 |
So the cpu is really high with a lot of dvol read.
Have we reached the limit of ns20 (I don't know how many dvol read can handle the ns20) ? How can I check what is doing the cpu ?
Thanks in advance
No Events found!
mark_rogov
14 Posts
0
March 14th, 2012 14:00
You need to look wider than just those numbers. The CPU at 93% is too high. Are you replicating any filesystems? Are you running a dedup process? Are you running a backup job? How many checkpoints are taken on this filesystem?
The 46MB dvol read is not necessarily too high, but does indicate a lot of reading
Hence my questions about what else is going on. In either case, with other jobs running at the same time, you might want to stagger them (e.g. don't run at the same time, but space them out a bit).
Also, take a look at other server_stats reporting: cifs-std, nfs-std. That might help you pinpoint a problem.
Best regards,
--MR
Rainer_EMC
4 Operator
•
8.6K Posts
0
March 14th, 2012 16:00
In my experience - FIRST check if you are running the latest OS version