Search indexing

In order to provide fast searching, JIRA creates an index of the text entered into issue fields. This index is stored on the file system, and updated whenever issue text is added or modified. It is sometimes necessary to regenerate this index manually; for instance when you add a new custom field, or if the index is lost or corrupted.

See Re-indexing after major configuration changes for more information on when you should re-index.

Note: For all of the following procedures, you must be logged in as a user with the JIRA Administrators global permission.

Re-indexing JIRA

  1. Choose > System
  2. Select Advanced > Indexing to open the Indexing page.
  3. This page allows you to choose one of the following two re-indexing options:

    • Background re-index — This will re-index all issues in the background.
    • Lock JIRA and rebuild index — This will delete and rebuild all indices, including the comment and change history indices.

Screenshot: Re-indexing JIRA

Which re-indexing option should I use?

The following table sumarizes the differences between both options:

Background re-indexLock JIRA and rebuild index
Single-threaded, slower to complete.Multi-threaded, faster to complete.

Can be canceled at any time.

Can't be canceled once started.
Keeps current index and updates it in-place.Rebuilds the index, optimizes it, and deletes the old one.
Causes disk fragmentation.Eliminates disk fragmentation.


Background re-index

This option allows any Jira instance to remain usable during re-indexing; however, the instance will be slower. If you have to perform this option, do so during a low-usage period. 

In addition, your disk fragmentation increases each time you perform a Background re-index. Whenever possible, perform a project re-index instead, especially if you just performed configuration changes that only affect one project. See Re-indexing a single project for instructions.

Lock JIRA and rebuild index

Use this option when the indexes are corrupt, which may be caused by a system or disk failure. This option deletes all indexes and rebuilds them, and is often called a full re-index

Overall, the Lock JIRA and rebuild index option provides greater benefits. The only downside for this option is that it'll lock a single-node instance, making it unavailable to users during the re-index.

On a multi-node Jira Data Center, you can use the Lock JIRA and rebuild index option without actually locking the instance. Therefore, if your Data Center instance has multiple nodes, don't bother with a Background re-index. See Re-indexing JIRA Data Center with no downtime for instructions.

Over time, your instance's disk becomes fragmented from normal use, which slows down your instance. Only the Lock JIRA and rebuild index option can address this. This means that, regardless of whether you're running a single- or multi-node instance, you should run a full re-index to address fragmentation periodically (for example, weekly or monthly).

You can also speed up a full re-index by increasing the number of threads it uses. See Tune number of index threads to make it go faster for details.

Instances with lower data complexity also perform full re-indexes faster. See Managing custom fields in Jira effectively for related information.


Choosing a custom Index Path

  • If you upgraded JIRA with an XML backup from a JIRA version prior to 4.2 and used a custom directory for your index path, you can choose between using this custom directory (which cannot be edited) or the default directory for your index path location. However, once you switch to using the default directory, you can no longer choose the custom directory option.
  • The default directory location is the caches/indexes subdirectory of the JIRA application home directory.

(info) NFS storage for JIRA indexes is not supported. See Supported platforms for more information.

Re-indexing Jira Data Center with no downtime

Keeping the integrity of indexes is as important as having your JIRA instance open to users all the time. These steps will help you run the Lock Jira and rebuild index option, which deletes and recreates all indexes, with no downtime.

Before you begin:

Choose a node and remove it from the load balancer. You'll use it to perform the re-index.

To re-index Jira Data Center with no downtime:

  1. Access JIRA on the node you've chosen, and select  > System.
  2. Select Advanced > Indexing to open the Indexing page. Then, run Lock Jira and re-build index.
  3. After the re-indexing is complete, take a look around the JIRA instance to make sure everything looks fine.
  4. Add the node back to the load balancer.

After completing the re-indexing, the rebuilt indexes will be automatically distributed to other nodes in the cluster (there might be some performance degradation during that time). If some changes were made to the indexes in the meantime, they will also be applied to maintain the integrity.

Backing up and recovering your index

Enabling index recovery will cause a snapshot of the indexes to be taken periodically. This allows you to recover your index quickly, rather than rebuilding the index, if there is a failure. This is particularly useful if you have a large JIRA installation and you cannot afford for it to be offline for long. If you have a small JIRA instance, it may not be worth enabling index recovery, as it rebuilding the index won't take much time.

Whether a full index rebuild is faster than recovering from a snapshot depends on a number of factors, including how recent the snapshot being recovered was taken. Large and complex installations should test this process on a development/testing server before relying on it in production.

To enable index recovery:

  1. Navigate to the Indexing page (as described above).
  2. Click Edit Settings to enable index recovery and choose the frequency of snapshots.
    • Snapshots are stored in the <yourjirahome>/exports/export/indexsnapshots directory.

To recover an index:

  1. Navigate to the Indexing page (as described above).
  2. Enter the name of the previously saved index in File name and click Recover.
    • JIRA will not be available during the recovery of the index. 
    • If changes were made to the configuration that required a re-index after the snapshot was taken, then you will need to do a background re-index after the recovery. Note, JIRA will be available after the recovery.

Additional information

  • JIRA will retain the last three snapshots at any time (in <yourjirahome>/exports/export/indexsnapshots). Older snapshots will be automatically deleted. Note, snapshots may occupy considerable disk space and may need to be moved to offline storage or deleted as appropriate.
  • The snapshot process is a relatively lightweight process and does not place much of a load on the system.
  • The process of taking a snapshot will require temporary disk space equivalent to the index size. The resulting snapshots will each be about 25% the size of the index.
  • All issues will be re-indexed appropriately during the recovery, including issues that were added, updated or deleted after the snapshot was taken.
  • You can use the index recovery process to bring your index up to date, if you need to restore your JIRA database. The index snapshot must pre-date the database backup being restored.

Re-indexing a single project

If you have made a configuration change that affects a single project, you can re-index just that project. See Re-indexing after major configuration changes for more information on when you should re-index.

To re-index a single project:

  1. Navigate to the desired project and click the Administration tab.
  2. Click Actions > Re-index project to start re-indexing the project.
Last modified on Oct 6, 2021

Was this helpful?

Yes
No
Provide feedback about this article
Powered by Confluence and Scroll Viewport.