Howto fix the River-of-News from now longer being updated

The River-of-News functionality in IBM Connections sometimes can get stuck.
All events that are happening inside the Connections platform are no longer being sent
to the Homepage Updates tab.

  When users puts up a new status for
  their Profile it will be visible on their
  own Profile page. But this event will
  not be pushed to the different
  sections on the Updates tab.

  For example the Status Updates
  section or the Discover section.

  For as far I’m aware all the other
  functionality in Connections will
  still work okay.

 

All will run fine until the messagestores get entirely clogged up and the queue reaches
the default max level of 50.000 messages. When this occur you will notify the system
becoming unstable.

Ofcourse the best approach is to fix it as soon as you notice that the River-of-News
no longer is being updated. Besides checking the Updates tab on the Homepage you
can check the following overviews in the WAS console to verify if messages are stuck.

( Screenshots are based on a medium Connections deployement. )

When the queue depths are around 70-100 and increasing you can be pretty
sure that  messages are stuck and are failing to be delivered.

Navigate to the Messaging Engines section and check if all of them are indeed
running else you can try to just start them up and see if that fixes it.

To fix clogged up message stores I found the quickest solution to stop all JVM servers
and delete all the messagestore files.

Find out the locations of the messagestore files by checking the File store section for
all the Messaging engines on your Connections system.

Delete all the files under this location and the one under the log directory. 
This action will only remove the events that are in the messagestore at
that moment. ( So starting from the point that messages were no longer
delivered to the River-of-News section )

Content that already reached the River-of-News will stay available as
messages that were received successfully are stored in the database.
The messagestores seem to be a temporary location to store event messages.


  For a medium deployment
  you wil have three sub
  directories to clean up.

 

After you have cleaned up the files in the above mentioned sub directories
you can start up the  JVM servers again.
Starting up the JVM servers will recreate these messagestore files.

Trying to stop, remove the messagestore files and startup the message
engines while the JVM servers were still up-and-running wasn’t a big success.

This entry was posted in ibm connections. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

Please leave these two fields as-is:

Protected by Invisible Defender. Showed 403 to 807,320 bad guys.