Jira Data Center (High Availability and Clustering)

As of v2.6.12 of the Git Integration for Jira add-on, the Data Center version of Jira is supported.  Jira 6.3 or higher is required for Data Center editions of Jira.

fyi
Known issue (admins): To provide access for both users for newly created tables, configure default tables and sequences permissions as follows (where jira1 & jira2 are users):
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL PRIVILEGES ON tables TO jira1;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL PRIVILEGES ON tables TO jira2;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL PRIVILEGES ON sequences TO jira1;
ALTER DEFAULT PRIVILEGES IN SCHEMA public GRANT ALL PRIVILEGES ON sequences TO jira2;

The configured repositories are located on the shared resource, since the nodes do not contain their own copy of the repositories.

fyi
The git repositories are located in the Jira Data Center shared folder.

The smart commits are processed by the reindex job which run once per cluster.  The node updates the index during the scheduled reindex job.  For Jira Data Centers, the indexes are stored in each node.

The GarbageCollection job should be run once on one node at the same time.  The following are indexing triggers:

  • Indexing is run on newly created repository on other nodes.
  • Entries from index will be removed from other nodes.
  • Reindex all repositories on other nodes when the reindex job is triggered.

For node changes:

When number of nodes in the Data Center cluster is changed, the Jira administrator should manually perform a full reindex for all repositories to rebuild index on new nodes.