Graph gaps - what does this mean?

agenjtoa · Post by **agenjtoa** » Tue Nov 15, 2011 12:51 pm

Hi,

I'm new with Cacti and reading my way through things however we also have Cacti which was setup by a previous colleague currently running.

One of the graphs that are printed are showing constant gaps (see screenshots). Not sure if this is due to connection problems or whether something to do with the SNMP data coming back - maybe a break or misconfiguration?

Thanks,

aT

Post by **TheWitness** » Tue Nov 15, 2011 10:46 pm

You may have an snmp timeout set too low. You need to review the log.

agenjtoa · Post by **agenjtoa** » Wed Nov 16, 2011 7:40 am

I've checked the system utilities (Cacti Log File, Poller Cache and SNMP Cache) and cannot see any indication that might suggest anything related to timeout. Anything specific that I need to be looking for other than timeout or anywhere else other than the mentioned above?

It is a Cisco 887, so I've changed the queue length to 20 and timeout to 60 and see how that goes.

EDIT: Still no joy after the SNMP timeout and queue change. Any ideas?

Post by **TheWitness** » Wed Nov 16, 2011 8:53 pm

Post your log over that time range.

agenjtoa · Post by **agenjtoa** » Thu Nov 17, 2011 8:36 am

Hi,

I've attached the logs.

aT

agenjtoa · Post by **agenjtoa** » Mon Nov 21, 2011 10:53 am

Further search and I found references to crontab being out of sync. I followed some of the forum posts on changing the poller interval to see what changes that may or may not show (poller.php was */1, set */5). However isn't plotting anything (I rebuild the poller cache and even bounced the box). Other devices seem OK, it seems to be just this one device. Deleted and recreated the graphs and still now joy.

Any other suggestions?

Post by **TheWitness** » Wed Nov 23, 2011 8:26 pm

SNMPv3 and is Cacti on a VM, or is the host a VM and is the time synchronized on both devices?

agenjtoa · Post by **agenjtoa** » Thu Nov 24, 2011 1:51 pm

There were 3 hosts setup for SNMP that weren't configured. Removed that and now at least SNMPv3 entries have cleared up.

Yes, Cacti sits on an ESXi box.

Time sync? Where do I verify if there is a time sync issue. Also, if this was the case, would it not affect all graphs? At the moment it seems only 1 host that is experiencing problem.

I've attached a list of updated logs, I'm not entirely sure if I'm made things worse by tampering with the crontab...

aT

Post by **TheWitness** » Thu Nov 24, 2011 2:08 pm

SNMPv3 will fail if the clocks are scewed up.

agenjtoa · Post by **agenjtoa** » Fri Nov 25, 2011 3:40 am

I'm getting our VMware people to install the VMTools to enable the time sync.

Post by **TheWitness** » Fri Nov 25, 2011 6:04 am

I'm just wondering you cause the dry humor in my response...

agenjtoa · Post by **agenjtoa** » Thu Dec 01, 2011 4:29 am

I think I may have missed that :S

I believe this is related to the VM platform issue, so for the time being, we'll have to work with it as is - everything else just breaks the graphs.

Cacti

Graph gaps - what does this mean?

Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Re: Graph gaps - what does this mean?

Who is online