[SOLVED] All Graphs Stopped At The Same Time

[alliedops]

I just inherited a production cacti installation when the fellow who built it suddenly didn't work here any more.

Previously I had used it, i.e., adding devices, creating graphs and adding users, but I know nothing about its innards.

It appears all graphs stopped updating about 4 days ago.

I have a screen shot here: http://openqbl.org/preview/cacti.jpg

Cacti 0.8.6j running on CentOS 2.6.9-42.0.10.plus.c4smp
I have root and am moderately fluent in *n*x, though most of my experience is on slackware boxes.

I will repeat I know nothing about cacti innards. I am slowly twisting in the wind and could really use a hero about now.

[fmangeant]

Hi

can you check that :

• the poller is running (check timestamp of log/cacti.log)
• your server time has not changed back (RRDtool does not like it)

[alliedops]

I found the immediate problem: the cacti.log file was 2.1 gigs, the limit for the file system. Not being able to write to the log apparently mucks things up. I set up a cron job to zero out the file periodically and graphing has resumed. Short term fix.

Now I need to find out why there are so many errors being written to the log. Here is a sample:

11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3959] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3959] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3961] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3961] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3962] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3962] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3963] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3963] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3964] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur
11/30/2007 02:46:47 PM - CMDPHP: Poller[0] Host[69] DS[3964] WARNING: Result from SNMP not valid. Partial Result: No Such Instance cur

Once I can figure out what is causing the errors I can try to address them.

[gandalf]

I found the immediate problem: the cacti.log file was 2.1 gigs, the limit for the file system. Not being able to write to the log apparently mucks things up. I set up a cron job to zero out the file periodically and graphing has resumed. Short term fix.

Please consider http://docs.cacti.net/node/435 as a long term fix.
Please re-open a new thread for your second problem
Reinhard

[alliedops]

I found the immediate problem: the cacti.log file was 2.1 gigs, the limit for the file system. Not being able to write to the log apparently mucks things up. I set up a cron job to zero out the file periodically and graphing has resumed. Short term fix.

Please consider http://docs.cacti.net/node/435 as a long term fix.
Please re-open a new thread for your second problem
Reinhard

Thank you, sir. I have an idea that the errors themselves are being created by discovered interfaces that either can't or aren't talking snmp. No all I've gotta do is figure out how that works. I love my job.