Active IQ Unified Manager Discussions

Performance Advisor: Not Enough Free Space

bowhuntr09
5,063 Views

Hi All,

We have been having an issue the past serveral days where Performace Advisor complains about being out of space. I changed the global option "eventsPurgeInterval" from 25.71 weeks to 13 weeks and that did not help. I have currently set the monMinFreePercent from 10% to 8.0% to temporarily stop the errors. My perfdata directory is currently at 28GB for 2 filers. I only have 37GB on the entire drive so I need to trim this data. Any ideas on how I can get this space issue under control?

1 ACCEPTED SOLUTION

kryan
5,062 Views

This file should be this large:

lun                       Netapp1                                  perf_8_72_2149            10080 1.25 GB         1.25 GB

However it is actually this large:

07/19/2011  10:16 AM     5,515,977,728 perf_8_72_2149

And there are multiple versions of this large file probably due to copy errors when space was low:

04/03/2011  12:29 AM     3,355,632,128 perf_8_72_1292

05/20/2011  12:15 AM     4,294,281,728 perf_8_72_1669

06/05/2011  12:30 AM     4,591,843,328 perf_8_72_1790

07/17/2011  12:34 AM     5,448,240,128 perf_8_72_2121

You can delete these 4 older files to gain back ~18Gb.

To correct the space discrepancy you will need to open a support case to get the work around for bug 439756.

View solution in original post

8 REPLIES 8

kirana
5,063 Views

hi.

Another approach i would suggest is to alter(relax) the Perf 'data collection' default intervals(collection interval, retention duration) - a feature introduced in DFM 4.0. To acheive this, in NMC PA-Setup-Hosts, select the host which is configured for Perf data collection, go to Edit-DAta collection, this would take you through wizard to alter the collection intervals and retention values for few/all of the CGs(counter groups). Especially volume CG would consume the highest space.

IIRC, the estimated space savings is displayed on the same page, on exitting this wizard also, for each CG.

kryan
5,063 Views

If you can post the output of "dfm perf data list -v" along with a matching perf data directory listing (as you already collected) we can determine if you are hitting bug 439756.

http://now.netapp.com/cgi-bin/bol?Type=Quick&bugs=439756

bowhuntr09
5,062 Views

Here is the output of "dfm perf data list -v"

kryan
5,062 Views

This bug only affects DFM 4.x - please post the dfm version as well.

bowhuntr09
5,062 Views

C:\>dfm about

Version                          4.0.2.7803 (4.0.2)

kryan
5,063 Views

This file should be this large:

lun                       Netapp1                                  perf_8_72_2149            10080 1.25 GB         1.25 GB

However it is actually this large:

07/19/2011  10:16 AM     5,515,977,728 perf_8_72_2149

And there are multiple versions of this large file probably due to copy errors when space was low:

04/03/2011  12:29 AM     3,355,632,128 perf_8_72_1292

05/20/2011  12:15 AM     4,294,281,728 perf_8_72_1669

06/05/2011  12:30 AM     4,591,843,328 perf_8_72_1790

07/17/2011  12:34 AM     5,448,240,128 perf_8_72_2121

You can delete these 4 older files to gain back ~18Gb.

To correct the space discrepancy you will need to open a support case to get the work around for bug 439756.

bowhuntr09
5,062 Views

Do I need to stop any services or anything before I delete those files or are they safe to just delete?

kryan
5,062 Views

These four files are older versions of the current file (2149) and should not be in use - you can simply delete them without stopping anything.

04/03/2011  12:29 AM     3,355,632,128 perf_8_72_1292

05/20/2011  12:15 AM     4,294,281,728 perf_8_72_1669

06/05/2011  12:30 AM     4,591,843,328 perf_8_72_1790

07/17/2011  12:34 AM     5,448,240,128 perf_8_72_2121

Public