(UCMDB) Support Tip: Discovery job stuck at 'Waiting for Probe' status error

1 Likes

Environment:
Any ucmdb versions.


Situation:
In the event the Discovery job gets stuck at "Waiting for probe" status for long time and does not proceed, and when try to get communication log the following error occurs:

'Cannot dispatch task on probe XXXX. Check if Probe Gateway is running'


Cause and Resolution:

This error is due to the Probe service not running, as a result the Probe is not connected to the UCMDB Server.
It is suggested to start Probe Services and check if Probe is connected to Server or not, in order to resolve such discrepancy.

Tags:

Labels:

Support Tip
Comment List
  • Here is another, more esoteric reason for trigger CIs appearing stuck for a long time.

    This time the trigger CIs appear hanging in "Reached Probe".

    This particular case I highlight can only hit you when you directly edit the discovery adapter pattern XML.

    I had changed the adapter XML, deleted a script:

                <script index="5">yet_another_library.py</script>

                <script index="3">another_library.py</script>

                <script index="2">a_library.py</script>

                <script index="1">main_script.py</script>

    But I forgot to edit the index attribute of the script tag -- you see, there is no script tag having index=4 anymore.

    UCMDB doesn't like this, but the error shows up later than expected, at the point of time when you want to run the discovery with a trigger CI.

    Your trigger CIs are appearing stuck after they reached the data flow probe (DFP) .

    In WrapperProbeGW Log you will find something like

    jvm 1    | <2022-11-03 19:35:32,291> 134489 [WARN ] [ProbeGW: DB Tasks Distributor] (TaskDispatcher.java:221) - Cannot retrieve job info - job won't be handled
    
    jvm 1    | java.lang.RuntimeException: Could not create DiscoveryJobInfo for job: MYDISCOVERYJOB
    
    jvm 1    |         at com.hp.ucmdb.discovery.probe.agents.probemgr.jobsmgr.DiscoveryJobsManager.getDiscoveryJobInfo(DiscoveryJobsManager.java:182)

    that when the DFP wants to process the job it fails.

    and a bit earlier in the exception you see the reason

    jvm 1    | <2021-11-03 19:35:32,291> 134489 [ERROR] [ProbeGW: DB Tasks Distributor] (DiscoveryJobsManager.java:483) - Error occur while trying to generate Discovery Job info for jobMYDISCOVERYJOB
    
    jvm 1    | java.lang.ArrayIndexOutOfBoundsException: 4
    
    jvm 1    |         at com.hp.ucmdb.discovery.probe.services.dynamic.core.BaseDynamicParams.parseXml(BaseDynamicParams.java:90)

    The missing index=4 caused an ArrayIndexOutOfBoundsException on the DFP

  • Some other causes/cleanups to mention:

    * correct outdated static entries in an adapters "Override default probe selection" field

    * review UriEndpointConfiguration.xml,, clean up configurations that reference a no longer existing probe) and run the "Manual UriEndpoint Discovery"

    * correct or delete UriEndoiunt CIs having URI Probe Name that reference a no longer existing probe

Related
Recommended