 | | | Rev 8 Seems to Break rrd [SOLVED] | Rev 8 Seems to Break rrd [SOLVED] 2006-07-22 - By Ben Mohilef
Back
I muttered:
> Anyone upgraded an athlon system (single cpu 32 bit) with munin and rrd
> running? The athlons seem to freeze (as in power down reboot) and a
> quick glimps of one running xsysinfo showed a heavy cpu load before fail.
> Logs are no help here. Intel (and even EPIA M10000) boxes don't exhibit
> the problem.
>
> Removing rrd and munin appears to resolve the issue.
>
> I seem to have seen a similar problem with new glibc's on athlons in the
> past. If that is true then other interesting things should start popping up
on
> athlon systems.
>
> Anyone have any ideas?
>
>
A upgrade of rrdtool to the 1.2 version wandered in somehow about the
same time. That version seems to need a higher version of python than is
on RHEL3, and munin ended up using massive resources whenever it
updated. The Intel and Epia boxes tolerated it, while the athlons ended up
in some kind of an infinite loop and died.
Reverting to 1.0.5 rrdtool version on the athlons fixed the problem.
Although the 1.2 and 1.5 versions do run on the Intel and EPIA boxes, the
1.0.5 version of rrdtool uses substantially less cpu time to execute.
I think that the newer versions are more appropriate for RHEL4, since this
problem is not apparent on an EL4 box.
--
Taroon-list mailing list
Taroon-list@(protected)
https://www.redhat.com/mailman/listinfo/taroon-list
|
|
|