perfSONAR User Q&A and Other Discussion

[Andrew Lake <>]

Hi Winnie,

You can delete those files. We are beta testing a new version of toolkit that 
will clean-up those files automatically. We are actively exploring the root 
cause of this build-up some people are seeing. Essentially testing is falling 
way behind for some reason and the problem compounds itself. I unfortunately 
don't have more details at the moment, but we're working on it. 

Thanks,
Andy




On Mar 5, 2015, at 7:05 AM, Winnie Lacesso 
<>
 wrote:

> 
> Good day Perfsonar People!
> 
> Our poor latency box got stuck around 22 Feb, its psGraph.cgi page became 
> blank & both disk space usage rose & especially %inodes used rose. Last 
> time %inodes used reached 100%, want to head it off at the pass this time!
> 
> regular_testing.log was timestamped 22Feb & looked full of errors.
> 
> Reboot ==> regular_testing appears to have revived ok but the disk space 
> usage & esp #inodes used are not reducing = this software not clean up 
> after itself v well :(   (I appreciate it's complex software......)
> 
> Having deleted most old ow* & bw* dirs in /var/lib/perfsonar/regular_testing
> from 5 Feb to 3 March there are STILL high %inodes used. 
> 
> The ow* & bw* dirs remaining in /var/lib/perfsonar/regular_testing
> are 4 March & 5 March, & ls -lR on them shows some ow* dirs have tens of 
> thousands of files in them (guess: that's the high %inode usage*)
> 
> To reclaim inodes, is it safe in /var/lib/perfsonar/regular_testing to do
> this :
> 
> service regular_testing stop
> /bin/rm -rf ow* bw*
> service regular_testing start 
> 
> ?  Don't want to personar testing state to get _worse_.
> 
> 
> Also: *are* tens of thousands of files in
> /var/lib/perfsonar/regular_testing/ow* directories normal?
> 
> If not, what might provoke this abnormality?
> 
> 
> 
> * I could be wrong about that, where else might pathologically high %inode 
> usage occur? 
>

Attachment: smime.p7s
Description: S/MIME cryptographic signature