>>> On 4/3/2008 at 4:30 PM, in message <[EMAIL PROTECTED]>, "Bernard Li" <[EMAIL PROTECTED]> wrote: > I noticed that since April 1st (that would be ~ r1185), I have been > getting the following message in my /var/log/messages: > > Apr 1 16:49:14 build01 /usr/sbin/gmetad[19901]: RRD_update > (/var/lib/ganglia/rrds/Build Servers/__SummaryInfo__/machine_t > ype.rrd): conversion of '^A' to float not complete: tail '^A' > > (some times I get ^B instead of ^A) > > Any ideas? >
I'm not sure why that is happening, but it brings up a bigger question. Why are we writing summaries for a string type? In the main loop of gmetad, it goes through a loop that does the root summary for the metrics. As part of this loop, sum_metric() is called which summarizes the numeric type metrics. However it is always doing a hash_insert even if the summary value didn't change (ie. string metric). Then in the next part of the main loop, write_root_summary attempts to write all metrics to RRD regardless of the value type. For string metrics, it is just constantly writing 0.0 to the RRD file. Is there any need to even create a summary RRD for a string type? We don't do it for host RRDs. Brad ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ Ganglia-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/ganglia-developers
