Hey Leigh:
Actually I just took a closer look at the error message, it seems that
your cluster nodes' time may not be synced properly... are you running
ntpd or something?
The error log seems to be saying that you cannot modify the database
because you are writing something in the past... sort of like Back In
the Future :-)
Cheers,
Bernard
Leigh Torgerson wrote:
Bernard,
Argh! I just checked again, and NOW it is working OK, but I haven't
changed anything. (well, shut down the machines for the night, started
them cold today and the reboot seems to have cleared up whatever
misgivings the rrd/gmond/gmetad combination was having..)
I had the machines running for a number of days, then the problem
appeared in the log files suddenly for two days - don't know what
precipitated it.
I've only had one piece of software running on the machines, but after
testing more today, I can't break it.. so it's a mystery. Guess I'll
write it off to a proton flare until I can reproduce the problem!
Thanks for responding,
Leigh
At 11:37 AM -0800 12/8/03, Bernard Li wrote:
Hi Leigh:
I wonder why that is happening - I am running RH9 and ganglia 2.5.5
and didn't get any of those error messages...
Cheers,
Bernard
Leigh Torgerson wrote:
We are running gmond/gmetad 2.5.5 on RH9 Linux machines - I noticed
that my /var/logs/messages files are filling up (multigigabytes -
glad I have a big HD :-) with error messages like:
Dec 5 16:27:39 ptl01 /usr/sbin/gmetad[4051]: RRD_update
(/var/lib/ganglia/rrds/__Su
mmaryInfo__/load_five.rrd): illegal attempt to update using time
1070670459 when las
t update time is 1070674058 (minimum one second step)
Dec 5 16:27:39 ptl01 /usr/sbin/gmetad[4051]: Unable to write meta
data for metric l
oad_five to RRD
Every time gmetad gets an update, errors like these are generated for
27 parameters. I saw from the list archives that this might have been
a bug in earlier releases that was hoped to be fixed in the latest
release. Apparently the fix didn't make it in..
Is there a way to turn this error reporting off or ?? I can probably
live with it by manually (or by cron) mucking with the messages file
every day to keep it from getting to be multigigabytes in size, but
maybe there's a better way!
thanks for any thoughts,
leigh
-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive? Does it
help you create better code? SHARE THE LOVE, and help us help
YOU! Click Here: http://sourceforge.net/donate/
_______________________________________________
Ganglia-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ganglia-general
