[Xymon] Max number of arguments to rrd?

[Henrik Størner]

On 17-12-2011 22:33, Tom Moore wrote:
> After the upgrade from 4.3,0 to 4.3.7, I seem to be having a strange
> issue where I have 48 datapoints I am trying to create the rrd from. It
> was working fine before, but now when my client sends the data it gives
> this error
>
> (names changed to protect the innocent)
>
> 2011-12-17 16:23:50 RRD error updating web.rrd from xx.xx.xx.xx:
> /usr/local/xymon/rrd/server.com/web.rrd: found extra data on update
> argument: 40:0:20:0:40:0:20:0:40:0:20:0:40:0:20:0
>
> I can’t seem to tell if this is coming from rrdtool or xymon. Based on
> the rrdtool docs, there shouldn’t be any sort of limit to the number of
> arguments. I have another exact monitor, that is only sending 2
> datapoints, and it is graphing fine. Just the 48 point one isn’t.
>
> Is there a way I can see what xymon_rrd is doing?

enable debugging on xymond_rrd, either by re-starting it with the 
"--debug" option, or do a "kill -USR2 <xymond_rrd processid>"

When calling the rrd-update function, it will log a message like:

Flushing '/jorn.hswn.dk/memory.actual.rrd' with 1 updates pending, 
template 'realmempct'

This gets translated into a command

   rrdtool update FILENAME -t TEMPLATE TIME1:VALUE1 <TIME2:VALUE2 ...>

In this case, TEMPLATE is 'realmempct' and there is "1 updates" - so 
there will only be one TIME:VALUE parameter.

Sending another USR2 signal turns off debugging.


It won't show the actual values, though, and I suspect that is where the 
problem stems from. This error usually means that there is a mismatch 
between the number of datasets listed in the template, and the actual 
number of data values present in the set of updates. Like if there were 
two memory-values in the update, but only one dataset- "realmempct" - 
listed in the template.

You can see the values in the error-message from the update, though. So 
comparing it to the template should be possible.

I am not sure, but another possibility is that the rrd file was created 
with a different number of datasets than what you are now trying to 
update it with. But I seem to recall that this would give another error.

Since web.rrd is a non-standard graph, I think you have a problem with 
the script generating these data.


Regards,
Henrik