How to rebuild AWI Index's

Products

• CA Automic Workload Automation - Automation Engine
• CA Automic One Automation

Issue/Introduction

When and how should I rebuild AWI's index?

You can rebuild indexes when there are problems in searching/listing newly created objects, or if you have no space to generate Lucene indexes. 

You may also witness that the search returns the objects which have been removed/deleted - In such cases rebuilding AWI's index help fix these.

Environment

Release : 12.x 21.x

Component : Automation Engine

Resolution

For V12.x

1. Stop the JCP's.
2. Remove the directory "indexCache" or "indexCache-12.X" from the bin directory of the Automation Engine.
3. Start the JCP's which will rebuild the index from scratch. The first started JCP on a node will take over this role.

For V21.x

1. Stop the JWP's
2. Remove the directory "indexCache" from the bin directory of the Automation Engine.
3. Start JWP's which will rebuild the index from scratch.

Note: On a distributed setup, this has to be done simultaneously on all the nodes.

Please also note that while the index is being (re)built, there could be some time where the search functionality does not work 100% accurately - this is to be expected.  How long it will take can be estimated using the below, but is not a for sure:

• The following line shows how many objects will be added to the indexcache:
  U00045136 Search index update: Detected 101192 objects which need to be re-indexed.
• The following lines will show how many indexes will be created
  U0045133 Search index update: executing 102 partition(s) for 101192 modified object(s).
• The following lines will show up after each partition is created and can be used to approximate how long each partition could take:
  U00045134 Search index update: partition #1 finished (6.116s).
• If you take the two above, you can determine about how long it will take: 102 partitions at around 6 seconds per partition == 612 seconds or about 10 minutes.

Additional Information

In version 12.x the JCP handled the rebuilding of the Lucene index.

With version 21.0 this index rebuild role is handled by the JWP. ( Ref. JWP Roles )

You can start a JWP as a dedicated process with a specific role. For example, you can use the following command to start the JWP as a dedicated process for the indexing (IDX) role:

java -Xmx2G -jar ucsrvjp.jar -Iucsrv.ini -role=IDX