Skip to main content

Hello

Our /var/lib/centreon-broker/ partition is getting full very fast the last couple of days.

There are lots of files like "central-broker-master_InfluxDB.queue.central-broker-master-InfluxDB###", actually 435 when writing this (yesterday less than 400 when we checked),

from

-rw-rw-r-- 1 centreon-broker centreon-broker  99999948 Nov 23 00:48 /var/lib/centreon-broker/central-broker-master_InfluxDB.queue.central-broker-master-InfluxDB267

to

-rw-rw-r-- 1 centreon-broker centreon-broker  14033256 Nov 28 14:07 /var/lib/centreon-broker/central-broker-master_InfluxDB.queue.central-broker-master-InfluxDB701

Restarted services this morning around 08:47 in the Central (centengine, gorgoned and cbd) and then in all pollers (centengine and gorgoned),

No errors found in central-module-master.log and central-rrd-master.log since then.


But we get lots of this in central-broker-master_InfluxDB.log, like every 15 seconds:

{"error":"partial write: unable to parse '4155,metric_id=319911,service_id=94541 time=1732319048,value=inf 1732319048': invalid boolean\nunable to parse '4155,metric_id=319933,service_id=94541 time=1732319048,value=inf 1732319048': invalid boolean\nunable to parse '4155,metric_id=319949,service_id=94541 time=1732319048,value=inf 1732319048': invalid boolean dropped=0"}


We have identified the host and service in Centreon and the check looks normal and working. I can see the timeline and graph of the last 31 days, everything seems OK.

 

We don't know if that error is related with the increasing number of files in /var/lib/centreon-broker/. What could be the problem?

 

We migrated the Central and pollers from CentOS to an Oracle Linux 8 host 2 weeks ago, running 22.10 and are programmed to update to 23.10 next week, but the partition will get full in a couple of days or less.

 

Regards,

Enrique

hello

I had a similar issue once when I was running 22.x as there was an incompatibility with the broker version on the poller, they were running a version more recent than the central due to some updates.

this was fixed with 23.04 where you can mismatching version, and there is no more issue when upgrading the platform.

be sure you are running a central version at the same version of the pollers until you upgrade on 23.x

there may be other problems however, to check version

rpm -qa | grep centreon 

and look the sub version of the broker on central and pollers

​@christophe.niel-ACT 

thanks for the reply. Checked that my pollers and Central have same components version.

 

With the help of support this have been solved.

  • Disabled perdata for the template used in the service with the errors.
  • Stopped cbd
  • Removed queue files
  • Restart cbd.

Its been ok for several days, partition used space is less than 2% and no more queue files.

 

I’ll enable perdata for the template after upgrading.

 

Regards,

Enrique

Reply


Terms & ConditionsCookie settings