HowTos/ManagingIndexesIn11.4.3

This HOWTO applies to customers who have upgraded to 11.4.3 and who have not yet rebuilt their indexes. In this scenario, indexing may not work after starting the new 11.4.3 system due to a change in index request record size.

Important

Due to some later improvements, the manual steps to manage the Indexer Service queue described below are not required in versions later than 11.4.3 - so this HowTo affects 11.4.3 customers ONLY.

Symptoms

1. The indexwork directory fills up and never drains
2. The SIS indexer log shows no activity

Procedure to fix

The following steps need to be executed as the root user.

1. Install and start 11.4.3

2. Stop the Indexer Service with:

 kill -15 $(omsetsvc -p indexer) 

3. Wait 2 or 3 minutes to ensure it has exited, then check the process table with:

 ps -ef | grep indexer

4. If the indexer process is still running, type:

 kill -9 $(omsetsvc -p indexer)

5. Remove the Indexer request queue with:

 rm /var/opt/scalix/??/s/indexwork/*

6. Restart the Indexer Service by typing:

 indexer

7. Check with ls -l that a pipe called notify appears in /var/opt/scalix/??/s/indexwork; this should look similar to:

 srwxrwxrwx 1 root scalix 0 Feb 11 17:34 notify

8. If it does not appear, stop/restart the indexer process and check again. If the problem persists, it is necessary at this point to restart the server with:

 omshut && omrc

9. Indexing should now work as expected

Next Steps

By this point, the symptoms described above should disappear. However, it is important to note that until user's indexes have been rebuilt, searching will still not be fully effective.