DX Application Performance Management

  • 1.  Can't parse errors in collector logs

    Posted Jan 08, 2010 09:15 AM
    After upgrade from 7.2.1 to 8.1.0 , we started seeing the below errors in collector logs. I assume these data files were corrupted.  1/08/10 04:40:23 PM GMT [ERROR] [Manager] Can't parse name of /collector_A/data/archive/temp.1256342400000.data as a timestamp; skipping.
    1/08/10 04:40:23 PM GMT [ERROR] [Manager] Can't parse name of /collector_A/data/archive/temp.1256688000000.data as a timestamp; skipping.
    1/08/10 04:40:23 PM GMT [ERROR] [Manager] Can't parse name of /collector_A/data/archive/temp.1252368000000.data as a timestamp; skipping.
    1/08/10 04:40:23 PM GMT [ERROR] [Manager] Can't parse name of /collector_A/data/archive/temp.1249430400000.data as a timestamp; skipping.  is there any way to fix these errors / datafiles onfly ?  ThanksPrasad  


  • 2.  Re: Can't parse errors in collector logs

    Posted Jan 08, 2010 09:25 AM
    We have started getting this error more frequently as well.   We followsed the information from the CA Wily KnowledgeBase. Answer ID 2195.    Symptom You notice the following messages in the Introscope Enterprise Manager log, where is specific to your system: 3/06/09 11:38:41 AM EST [ERROR] [Manager] Error executing task com.wily.introscope.server.enterprise.entity.fsdb.ReperiodizeTask: java.lang.ArrayIndexOutOfBoundsException:
    ..
    3/06/09 11:58:57 AM EST [ERROR] [Manager] Can't parse name of data\archive\temp. .data as a timestamp; skipping.
    3/06/09 11:58:57 AM EST [ERROR] [Manager] Can't parse name of data\archive\temp. .data as a timestamp; skipping.
    3/06/09 11:58:57 AM EST [ERROR] [Manager] Can't parse name of data\archive\temp. .data as a timestamp; skipping.
    ...  Resolution These error messages  indicate that the EM was not shut down gracefully and some of the SmartStor files are corrupt. To solve the problem:1.  Stop the EM.2. Backup and delete  the  data files mentioned in the log file. For example, if the log  says: 3/06/09 11:58:57 AM EST [ERROR] [Manager] Can't parse name of data\archive\temp.1233896400000.data as a timestamp; skipping.
    3/06/09 11:58:57 AM EST [ERROR] [Manager] Can't parse name of data\archive\temp.1233982800000.data as a timestamp; skipping. Then you need to delete data\archive\temp.1233896400000.data and data\archive\temp.1233982800000.data files.3.  Start the EM.After  applying these  changes,  run tests  to see  if the problem is still occurring. If it is, contact CA Wily Support and provide  the contents (in a  ZIP file)  of the Introscope Enterprise Manager /logs directory. Is there a particular reason WHY this is happening?
       


  • 3.  Re: Can't parse errors in collector logs

    Posted Jan 08, 2010 09:37 AM
    we noticed these symptoms several times and reported to tech support. they suggested to remove those file as well . we tried that option but that doesn't solve our problem as it complains about missing references and data files after EM / collector restart.  ThanksPrasad


  • 4.  Re: Can't parse errors in collector logs

    Broadcom Employee
    Posted Jan 08, 2010 12:06 PM
    Hi Prasad,  The above errors indicate that the "reperiodization" process did not finish successfuly and these  temp files of the metadata were not deleted as expected.Usually you should see the following messages:12/14/09 11:33:12 AM PDT [VERBOSE] [Manager.SmartStor] Startup reperiodization of tier 1 took 0 seconds
    12/14/09 11:33:12 AM PDT [VERBOSE] [Manager.SmartStor] Evaluating data tiers...
    12/14/09 11:33:12 AM PDT [VERBOSE] [Manager.SmartStor] Evaluate tiering took 0 seconds
    12/14/09 11:33:12 AM PDT [VERBOSE] [Manager.SmartStor] Reperiodize took 0 seconds
    12/14/09 11:33:12 AM PDT [VERBOSE] [Manager.SmartStor] Reperiodize took 0 secondsFor now, you can safely remove these files.Reperiodization is not dependent on these temp files. They are created in the process of taking a backup of the metadata file and for some reason they did not get deleted.Please note that Reperiodization kicks off everytime the EM is restarted.  Now, based on your latest comment, I believe we need to review your configuration in detail again, so I would suggest you to file a ticket with support and make sure to attach the zipped content of the EM_HOME/logs directories from all collectors, IntroscopeEnterpriseManager.properties, hardware details (#cpu, cpu spped, memory) and screenshot of the following supportability metrics:  1) EM | Smartstor | Metadata | Metrics with Data2) EM | Internal | Number of Connection Tickets3) EM | Internal | Number of Virtual Metrics4) EM | Tasks | Harvest Duration5) EM | Tasks | Smartstor Duration6) EM | GC Heap | Bytes in Use7) EM | GC Heap | GC Duration  Make sure to enable VERBOSE logging level and  also add  to the incident a  sample of one of your IntroscopeAgent.profile and log file.  This is the minimum information we need to have a better understanding of the issue.It could be the case that the EMs are overloaded or  a metric explosion on the EM side might be causing these issues.  I hope this helps,  Regards,  Sergio