We have created a custom discovery pattern to parse XML files and create custom CI in uCMDB. We need to create around 50K CIs in uCMDB. DDM is processing the CI creation at a very slow pace (approx. 100 CIs/minute). From WrapperProbeGw.log, i could see the discovery script completes the processing of the XML file. But the probeMgr-result-processing.log shows that the discovered results are processed and CIs are created and stored in the probeMgr database at a slow pace.
My question is how can we get this processing speed tweaked and increased to process hundreds of thousands of CIs.
Thanks in advance, Mugilan
P.S. This thread has been moved from Application Perf Mgmt (BAC / BSM) Support and News Forum to CMS and Discovery Support and News Forum. -HP Forum Moderator
Thanks for replying to my query. I checked the discovery probe JMX console "type=TaskResultsDistributer ". The "NumberOfCurrentUnsentResults" is always showing as 0. But the results are being processed by the probe. In the "probemgr-result-processing.log" I could see several lines like the following one.
<2010-05-06 01:46:20,961> [DEBUG] [JobExecuterWorker-0:Ala Elements_188.8.131.52] (ResultProcessRedundant.java:185) - Storing Result Record in Probe database - Result will be Reported to Server <2010-05-06 01:46:20,961> [DEBUG] [JobExecuterWorker-0:Ala Elements_184.108.40.206] (ResultProcessRedundant.java:151) - jobID:Ala Elements,className:contains,triggerCiID:6bbec68236f9579efdbd990b8c420e01,cmdbObjID:4381735f2e4bb75450e2221d0a51b48f,objHashID:87f11250a1d2e260e63f3c372567e79c
In the JMX, "type=TaskDispatcher" - "getSummaryReport" shows the summary of the CIs stored in the probe database yet to be sent to CMDB (several thousands actually). Not sure why "NumberOfCurrentUnsentResults" is showing as 0.
I didn't try the "flushProbeResults" yet. I am waiting for the current discovery job to finish so that I can get the data in CMDB. Once the job is finished, I will run discovery with much smaller data set and try the flushresults. I will update the findings here.
Thanks again for responding to this query and appreciate your inputs to resolve the issue.