Unexpected user cache refresh in Jira causing various issues (failed login, missing users)

Still need help?

The Atlassian Community is here for you.

Ask the community


Platform notice: Server and Data Center only. This article only applies to Atlassian products on the Server and Data Center platforms.

Support for Server* products ended on February 15th 2024. If you are running a Server product, you can visit the Atlassian Server end of support announcement to review your migration options.

*Except Fisheye and Crucible

    

Summary

The Jira user cache gets unexpectedly refreshed, and as a result various symptoms might be observed on specific Jira nodes (in a Data Cluster environment):

  • some users might fail to login (depending on which nodes they were redirected to)
  • some users might be missing from the page ⚙ > User Management > Users
  • if a Confluence application is configured with the Jira application via an application link to sync users from Jira, some users might be missing in Confluence and fail to login

Environment

Jira Data Center on any version from 8.0.0

Configuration Manager for Jira before version 6.13.3

Diagnosis

Check the Jira logs from each node and look for logs showing the messages "clearing cache state" or "removeAll called on a pinned cache" triggered by the end point /rest/configuration-manager/1.0/cacherefresh:

2023-02-06 11:55:18,967 http-nio-8009-exec-25 INFO test 715x2285109x1 9fxipi /rest/configuration-manager/1.0/cacherefresh [c.atlassian.agile.support] 
    
    *********************************************************************************
    Atlassian GreenHopper v8.20.10 #b92c120b3532fd60 built 2022-06-22T16:16:50.081+02:00 - onClearCache() - clearing cache state
    *********************************************************************************
2023-02-06 11:55:19,186 RMI TCP Connection(5397)-10.24.220.20 WARN      [n.sf.ehcache.Cache] Data availability impacted:
    ****************************************************************************************
    ************************** removeAll called on a pinned cache **************************
    ****************************************************************************************


Shorter after this message, check if you see errors showing that some users don't exist:

2023-02-06 11:55:19,396 http-nio-8009-exec-7 ERROR anonymous 715x902287x1 10yk4qm  /download/resources/com.tempoplugin.tempo-core/img/io/my_work_onboarding/03_onboarding_action.png [d.c.t.integration.atlassian.PlJiraAuthenticator] User does not exist.
2023-02-06 11:55:19,397 http-nio-8009-exec-7 ERROR anonymous 715x902287x1 10yk4qm  /download/resources/com.tempoplugin.tempo-core/img/io/my_work_onboarding/03_onboarding_action.png [d.c.t.integration.atlassian.PlJiraAuthenticator] User does not exist.


If you verified that these cache refresh messages were triggered by the endpoint /rest/configuration-manager/1.0/cacherefresh ((warning) and not any other end point), then this KB article is relevant.

Cause

The endpoint /rest/configuration-manager/1.0/cacherefresh belongs to 3rd party add-on Configuration Manager for Jira (CMJ). Versions of CMJ before 6.13.3 had the option to clear caches or send a clear cache event.

Solution

The immediate solution consists of re-starting the impacted Jira nodes (nodes where you found that the add-on triggered a user cache refresh).

As for the long-term solution, we recommend upgrading to CMJ 6.13.3 or if you’re currently using an older version, DO NOT USE the ClearCacheEvent option or the Clear Caches button.

Last modified on Dec 4, 2023

Was this helpful?

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