HPE_PMDB_Platform_TaskManager service status shows wrong status

  • KM03632525
  • 06-Apr-2020
  • 07-Apr-2020

Summary

HPE_PMDB_Platform_TaskManager service status shows wrong status

Error

On some systems TaskManager service is running, but we observe below:

TaskManager service shows Deactive with systemctl:
# systemctl status HPE_PMDB_Platform_TaskManager.service
* HPE_PMDB_Platform_TaskManager.service - HPE Operation Bridge Reporter TaskManager
   Loaded: loaded (/usr/lib/systemd/system/HPE_PMDB_Platform_TaskManager.service; enabled; vendor preset: disabled)
   Active: deactivating (stop-sigterm) since Tue 2020-03-31 01:36:32 EDT; 2 days ago
  Process: 12211 ExecStop=/opt/HP/BSM/PMDB/bin/HPE_PMDB_Platform_TaskManager stop (code=exited, status=0/SUCCESS)
  Process: 18570 ExecStart=/opt/HP/BSM/PMDB/bin//HPE_PMDB_Platform_TaskManager start (code=exited, status=0/SUCCESS)
 Main PID: 18591 (code=exited, status=0/SUCCESS)
   CGroup: /system.slice/HPE_PMDB_Platform_TaskManager.service

Cause

When the server is not shutdown cleanly, it can leave zombire processes or not enough memory for the service to run.

Using the SIGTERM command can cause the command-line to hang.

Fix

  1. Ensure Flink is configured correctly according Sizing guide and has appropriate values based of OBR distribution scenario and resources of server
  2. Check that the /opt/HP/BSM/PMDB/bin/HPE_PMDB_Platform_TaskManager.service has the correct values.
  3. Verify /usr/lib/systemd/system/HPE_PMDB_Platform_TaskManager.service has correct permissions.  In some scenarios not having executable permissions can cause this issue.
  4. Stop and Start the service as well as check /var/log/messages for TaskManager events.
  5. Linux Zombie processes can be related to Active: deactivating (stop-sigterm) or not a clean shut down of the server. Rebooting the server could resolve this issue.