• Icon: Suggestion Suggestion
    • Resolution: Unresolved
    • None
    • None
    • 107
    • 2
    • We collect Jira feedback from various sources, and we evaluate what we've collected when planning our product roadmap. To understand how this piece of feedback will be reviewed, see our Implementation of New Features Policy.

      Currently the only way to clear all of Jira's cache is to restart.

      This request is for the ability to clear Jira's internal caches in the admin section, much like Confluence's cache management feature

            [JRASERVER-66095] Ability for Jira admins to flush internal caches

            rad added a comment - - edited

            yes

            rad added a comment - - edited yes

            Let me expand further on this:

            "Currently the only way to clear all of Jira's cache is to restart."

            That might have been true for Jira Server, but for Data Center correct me wrong, this trick does not entirely work now does it? And even if, are we suppose to restart nodes multiple times a day when our support team changes something in Jira and nodes(all except the one where it was changed) get caches out of sync?

            Let's not forget restarting one Jira node is question of 5+ minutes, and re-balancing of users is also fun to watch since yes, it does impact performance seen by end users during said procedure...

             

            Because if your solution is to restart whole cluster, incur downtime, then I would remove that HA part from description for DC.

            So that this does not become false advertising.

            This is bug alright if the above is true. Get it right.

            Tomáš Karas added a comment - Let me expand further on this: "Currently the only way to clear all of Jira's cache is to restart." That might have been true for Jira Server, but for Data Center correct me wrong, this trick does not entirely work now does it? And even if, are we suppose to restart nodes multiple times a day when our support team changes something in Jira and nodes(all except the one where it was changed) get caches out of sync? Let's not forget restarting one Jira node is question of 5+ minutes, and re-balancing of users is also fun to watch since yes, it does impact performance seen by end users during said procedure...   Because if your solution is to restart whole cluster, incur downtime, then I would remove that HA part from description for DC. So that this does not become false advertising. This is bug alright if the above is true. Get it right.

            Yeah, please do NOT use this Scriptrunner "Clear Jira Caches" in a clustered environment. It caused me a large outage due to the User Cache not being recreated on 3/4 of the application nodes.

            If you need to clear caches, just take the node out of the Loadbalancer, let it restart, let it catch up on indexes, then add it back to Loadbalancer.

            Steven Behnke added a comment - Yeah, please do NOT use this Scriptrunner "Clear Jira Caches" in a clustered environment. It caused me a large outage due to the User Cache not being recreated on 3/4 of the application nodes. If you need to clear caches, just take the node out of the Loadbalancer, let it restart, let it catch up on indexes, then add it back to Loadbalancer.

            Tomas Karas added a comment - - edited

            About ScriptRunner "built-in script" that is incorrectly tagged as "safe to run in production" is far from safe on data center.
            It logs out everyone, no-one can login, and the process can take tens of minutes,..... at which point we were doing restoration.....
            Not to mention in our case it dropped a bomb and DB index for users was damaged - we had to reindex them to restore the tool, otherwise the groups of users was complete and utter mess. Jira was unable to cope with that.

            Cuz that is clearly not available tool if you ask me.
            It does not say this anywhere in documentation, and we've run it assuming it clears cache of one node, not whole cluster at once.

            Tomas Karas added a comment - - edited About ScriptRunner "built-in script" that is incorrectly tagged as "safe to run in production" is far from safe on data center . It logs out everyone, no-one can login, and the process can take tens of minutes,..... at which point we were doing restoration..... Not to mention in our case it dropped a bomb and DB index for users was damaged - we had to reindex them to restore the tool, otherwise the groups of users was complete and utter mess. Jira was unable to cope with that. Cuz that is clearly not available tool if you ask me. It does not say this anywhere in documentation, and we've run it assuming it clears cache of one node, not whole cluster at once.

            This should have been in the core product out of the box.
            Especially in case of Data Center.

            Tomas Karas added a comment - This should have been in the core product out of the box. Especially in case of Data Center.

            MattS added a comment -

            This would be useful. I believe that ScriptRunner has a built-in script to do this

            MattS added a comment - This would be useful. I believe that ScriptRunner has a built-in script to do this

              Unassigned Unassigned
              jcurry Jeff Curry
              Votes:
              150 Vote for this issue
              Watchers:
              92 Start watching this issue

                Created:
                Updated: