Halp! All My Graphs have Stopped!

[lmiltchev]

Run the following command:

Code: Select all

rm -rf /usr/local/nagios/var/spool/perfdata/*
watch 'ls /usr/local/nagios/var/spool/perfdata | wc -l'

watch it for a while and see if the number of files is constantly growing. Is the npcd running?

Code: Select all

service npcd status

On the mrtg config verification - run the command without the "-check":

Code: Select all

LANG=C LC_ALL=C /usr/bin/mrtg /etc/mrtg/mrtg.cfg

As you add more and more devices, it is perfectly normal for it to take longer and longer, as you might expect. However, if it takes more than 5 minutes to complete a run, and if there are a lot of timeout messages, a cleanup is a very good idea. Are you seeing either a longer than 5 min runtime or lots of timeouts when running manually?

[BenGatewood]

OK, watching the perfdata spool it is hovering between 10 and 20 now.

NPCD is running

LANG=C LC_ALL=C /usr/bin/mrtg /etc/mrtg/mrtg.cfg returns an error concerning an offline host which I have now deactivated and then stacks of "did not eval into defined data" errors

--------------------------------------------------

Perfdata spool is creeping up - between 30 and 40 now.

More importantly, every single one of my interface bandwidth checks is still reporting 0 in and out when I know this is not the case.

[BenGatewood]

OK, think I've cleared it. BW values starting to appear correctly. Will continue to clean up the mrtg config files and see how it goes. Thanks for all you're help so far. I'll update this thread tomorrow to confirm status.

[slansing]

MRTG/rrdtraff checks can take a bit of time to come through when first added. This is because they need a base level data set form 2-3 or so checks first.

[slansing]

Was this only pertaining to your bandwith graphs? Your opening post seemed to indicate that "all" graphs were not being populated.

[BenGatewood]

Sorry - the original problem was *no* graphs were populating but I fixed that. It was then that I noticed that the BW graphs still weren't working and I found that every single BW check was reporting a value of 0 in and 0 out. This is what appeared to be working when I left last night after cleaning up the mrtg config files but when I've come back in this morning I have found they stopped working again a couple of hours later.

And still not working. Bandwidth graphs worked briefly this morning after I manually ran mrtg but then immediately stopped again. Running mrtg returns no errors now (with or without -check) and all other graphs are working - just bandwidth is broken.

I really don't know what's going on. It seems that BW checking works briefly and then stops again. All my BW graphs from the last 24 hours look like this:

BW.jpg

[sreinhardt]

My guess is that mrtg is still having issues with the number of ports\devices you are checking, could you run the following command and PM the resulting file in /tmp to one of us, don't post it on the board since it will have snmp community string and internal IP information.

Code: Select all

cd /tmp
tar cjf mrtg.tar.bz /etc/mrtg/

[BenGatewood]

Any luck with the files, guys?

[lmiltchev]

We are looking at this and will let you know what we found out sometime today.

[BenGatewood]

Thank-you. I'll be available all evening (GMT) if you need further info at all.