[rrd-users] Slow collection runtimes occurring regularly

[Steve Shipway]

> Thanks for the ideas but no, we're not using MRTG. rrdcached is under
> testing but not currently an option.

The RRD structure looks just like an MRTG-generated one!  Ah well, bad guess. Shame rrdcached is not an option yet (the trunk version supports many more functions) as it is a definite win on IO-bound performance.

While I'm not an expert on the internal programming of the RRDTool calculation mechanism, I would assume that (without rrdcached) the consolidation calculations are done at the time that the rrdupdate function is called, if the update time mandates it.  So, spreading your update calls over the 5min step window should spread your CPU loading over this time too (use an explicit timestamp rather than 'N' to preserve your data correctness, as Simon points out).  This approach works with MRTG/RRD but since I don’t know what collection mechanism you're using, YMMV 

Steve

Steve Shipway
ITS Unix Services Design Lead
University of Auckland, New Zealand
Floor 1, 58 Symonds Street, Auckland
Phone: +64 (0)9 3737599 ext 86487
DDI: +64 (0)9 924 6487
Mobile: +64 (0)21 753 189
Email: s.shipway at auckland.ac.nz
 Please consider the environment before printing this e-mail 


-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 4928 bytes
Desc: not available
Url : http://lists.oetiker.ch/pipermail/rrd-users/attachments/20110502/b00a2ab0/attachment.bin