Poller hangs on second normal poll

Document ID : KB000023359
Last Modified Date : 14/02/2018
Show Technical Document Details

Question: 

The eHealth server has been down for a considerable period of time; what would cause this to happen?

Environment:  

eHealth (all supported versions)
All supported platforms

Answer: 

On the second poll after a server restart, the poller performs availability backfill which can potentially take several hours depending on the server specifications and the size of the configuration. The server will calculate availability for all elements for 1800-second intervals for up to 72 hours.

To determine whether the availability backfill is taking place, monitor the size of the $NH_HOME/tmp/STSample files.  If these files are increasing in size and the time of these files is incrementing, the poller is functioning normally and is not in a hung state. 

If the poller is not allowed to finish this on the second poll because the server is restarted, the process will start over again so the servers should not be interrupted during this process. A message will appear in the system log and in OneClick to notify users that backfill is taking place.  You will not see any new statistics tables created until the backfill has finished. Once the backfill is complete the STATS0 tables will show 2 rows per element per table.