-reindex

Repotools command that regenerates triple store queries and Lucene text store indexes.

Purpose

The repotools reindex command is used against an offline server and affects all resources in the storage services. By default, the query triple store and the Lucene text store are regenerated, but the "scope" parameter can be used to rebuild either the query or the text indexes or both.

Re-indexing in a cluster environment

Single node re-indexing

The repotools reindex command can re-index a stand-alone Jazz server, or a server node in a cluster environment. For the latter, the server must be temporary isolated from the cluster for the maintenance duration. To accomplish this, you must first disable the server in the load balancer to avoid request being dispatched to this node during the shut down operation, and then stop the server.

Full cluster re-indexing

To perform a full cluster re-indexing, the reindex command must be issued on an isolated node as explained above. Using the "cluster" parameter, the command is then propagated to all the nodes in the cluster. The running servers perform the operation immediately depending on the state of its indexing service (activated or not), or like any other stopped server during the next start up.

The following example rebuilds the text indexes for all the nodes:
repotools -reindex scope=search cluster
While this example only rebuilds the text indexes on the node where the command is issued:
repotools -reindex scope=search

Re-indexing an online server

Use this method if you must keep your server online during re-indexing. Re-indexing a cluster this way can be a time consuming operation. Here are the basics steps to re-index a server from cloned indexes:
  1. Isolate and re-index a server from the cluster.
  2. Track the indexing operation for completion.
  3. Backup the server and deploy to the selected nodes, after proper isolation from the cluster.
  4. Restart the cloned server.
Note: When the server is up and running, some resources might be added to the indexes during the cloning process, but each cloned server will fix up the additional indexes using a standard consistency checks during the next restart.

Tracking re-indexing progress

For the running servers, the re-indexing progress can be tracked by accessing https://{JazzServerURL}/jazz/indexing, where {JazzServerURL} is either a single node URL that tracks a specific server, or the cluster URL that aggregates all the nodes re-indexing activities into one result.

Parameters

Attribute Description Required Default
teamserver.properties Path to the teamserver.properties file. Yes conf/jazz/teamserver.properties
scope The scope of reindexing. Values are all, query, and search. Yes all
cluster Reindex all nodes in the cluster. No  
contextPath Repository context path. No /jazz
noPrompt Do not prompt before reindexing a cluster. No
baseline A comma separated list of baselines URI, without spaces. No none

Example

repotools -reindex teamserver.properties=conf/jazz/teamserver.properties scope=all

Feedback