problem with queues on spooler probe

[davidms]

Hi

I have a problem with some nimsoft agents, the agente is not inserting information to uim DB, we review the problem and we saw that on the agent is withholding the information on the queues of spooler , the probe spooler has many expired messages and we see that the files q1.rdb and q2.rdb they are growing very fast, we increase space in the filesystem where the agent is installed but the files q1.rdb and q2.rdb they continue growing very fast. we reinstalling the agent but the problem persist.

Could you help us please?

We attach some screen

Best regard.

[Gene Howard]

Set your controller loglevel to 3 and logsize to 35000 and check for error communicating to the hub.
Also, change the bulk size in the spooler to 100 or 500.
By default, the robot spooler sends one message at a time.
if you have a robot that is collecting a lot of qos or generating a lot of alarms this could be too low.,

Also, make sure your robots and hubs are both on current versions 7.97 or higher.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[davidms]

Hi Gene

We applied the recomendations that you indcated us but the problem continue, we attach some screen abot the spooler probe configuration.

Also the robot is collecting a lot of qos (Metrics), do you beleve that we needs increasig the values on the spooler configuration?

The Robot versions is 7.97 HF9

Best regard.

[David Michel]

What is the linux distro and version? 

'a lot of qos (Metrics)' is rather subjective, so it might help to know in more details things like the monitoring probes, number of profiles, internal, etc. 


------------------------------
Support Engineer
Broadcom
------------------------------

[David Michel]

"The OS Version weher is installed the robot is Soalris 11, we have monitoring OS wit CDM, Oracle data base  with 4 profiles (every one with 15 metrics includs the SGA metric)."

Soaris 11 should be fine, and can't imagine just that for probes would overload it. 
If not already get the two probes at the GA build. 
Check the OS system logs. 

What's the history on this? 
new deployment and always had the problem
been running fine for some time then all of a sudden... 
or whateve 

Is there an abnormal usage of cpu, memory, or ports going along with this? 

As a test deactivate cdm, and restart the robot service. 

Also as Gene said got to check the spooler and controller logs. 

------------------------------
Support Engineer
Broadcom
------------------------------

[davidms]

Hi David

This monitored server was working very well for many time, but suddenly it started having problems,  the files q1.rdb and q2.rdb they starting to growing very fast, we increase space in the filesystem where the agent is installed but the files q1.rdb and q2.rdb they contined to growing very fast. we reinstalled the agent but the problem persist.

We stoped the cdm probe and after we restarted the robot, the problem cotinue.

We attach some screen about the spooler and controller logs, and the use of resources of CPU and memory.

Best regard.

[Gene Howard]

stop the robot.
Clear the niscache folder
remove the *.rdb files
restart and see if that fixes it.
Make sure ALL probes on the system are on the latest version.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------

[davidms]

Hi David

We were able to solve the problem, thank you very mucho for the help on the problem.

Best regard.

[Gene Howard]

you would need to check the spooler logs and controller logs for errors
I do not think I have ever seen anyone need to put this above 500.
Make sure you do not have antivirus scanning the robot directory.
You can try removing the spooler *.db files from the robot directory and let them recreate.

------------------------------
Gene Howard
Principal Support Engineer
Broadcom
------------------------------