Cacti graph gap

[homi_mofo]

Hy everybody,
I have a small problem with cacti that drives me crazy, sometimes and just sometimes the graph is emty, after a while it recovers by it self, and graphs again. It is annoying, it does this not for just 1 equipement but for all, ()diffrent from manufacturer)
I have tweaked the SNMP Timeout, SNMP Retries from General and Script and Script Server Timeout Value. And still no luck.

cacti.log file
03/13/2009 02:00:06 PM - SYSTEM STATS: Time:3.2478 Method:cmd.php Processes:1 Threads:N/A Hosts:4 HostsPerProcess:4 DataSources:47 RRDsProcessed:26
03/13/2009 01:55:05 PM - SYSTEM STATS: Time:3.4648 Method:cmd.php Processes:1 Threads:N/A Hosts:4 HostsPerProcess:4 DataSources:47 RRDsProcessed:26
03/13/2009 01:45:06 PM - SYSTEM STATS: Time:3.2464 Method:cmd.php Processes:1 Threads:N/A Hosts:4 HostsPerProcess:4 DataSources:47 RRDsProcessed:26
03/13/2009 01:40:05 PM - SYSTEM STATS: Time:2.2634 Method:cmd.php Processes:1 Threads:N/A Hosts:4 HostsPerProcess:4 DataSources:47 RRDsProcessed:26


what could it be?[/img]

[TheWitness]

Something to do with Cron. Are you running on a VM? If you cron the poller with the "-f" option, it will always run.

TheWitness

[mknnoc]

Hi, can you post the actual config on cron file?

[TheWitness]

Look at the table poller_item and the field rrd_next_step. Insure that it is a signed integer. Then reset it's values all to 0. Here's some sql.

Code: Select all

ALTER TABLE `poller_item` MODIFY COLUMN `rrd_next_step` MEDIUMINT(8) NOT NULL DEFAULT 0;
UPDATE poller_item SET rrd_next_step=0;

TheWitness

[mknnoc]

so, what is the actual problem?

[TheWitness]

I'm guessing mind you, that you are either on 0.8.7a/b and have not upgraded to 0.8.7c/d. In the earlier version, the field was unsigned, and the poller relies on the next step going negative. When its unsigned, instead of going negative, it underflows and get's real big. Then, it has to count down. So, it does not poll again for a while.

This may not be your problem however. You have to help me as I have no access to your system.

TheWitness

[mknnoc]

Hi Witness,

i can let you login to the system in order to investigate the problem. and give out the bug fix. Really appreciate your helps!!

pls let me know your email or IM.

I am using 0.8.7d version with centos5.2

[CaliMale]

I am running the same version of cacti, same version of CentOS and I have the exact same problem.

Did you find a solution? Can share the love?

Thanks!

[colkinz]

I am also experiencing exactly the same issue with 0.8.7d.

The database was updated from 0.8.7b but looks fine:
`rrd_next_step` mediumint(8) NOT NULL default '0'.

[star3am]

Hi Witness, I'm having a small problem with gaps too ..

I have tweaked mysql and php and the OS, but nothing that made me sit up yet,

you mentioned,

Are you running on a VM?

I am running cacti on a VM, but the machine looks fine,



Xen VM
Centos: 5.3
Cacti: 0.8.7d
Spine Poller:

05/24/2009 02:19:29 AM - SYSTEM STATS: Time:27.9578 Method:spine Processes:50 Threads:15 Hosts:18 HostsPerProcess:1 DataSources:492 RRDsProcessed:227

Thank you for the help

ciao/Riaan

[Thomas.Pacce]

Same problem here, i run Cacti Version 0.8.7b on a VM 9ubuntu server 8.10).
I see gaps on most of my hosts monitored for a couple of days now.

[star3am]

anyone ?

[dieselboy]

i should have read before posting a new topic, i am having this exact same issue.

i am running 0.8.7d, under PPC ubuntu 7.10

[star3am]

Hi, just an update ..

think it's safe to rule out RRDtool, as I have upgraded to various versions, from 1.2.7 - 1.3.7 which I am on currently ...

[dieselboy]

what are the configurations in your crontab ?