Confluence 7.17.1 crashes by transaction PluginReadWriteTx

w_v_krimpen_slootweg

Server runs out of memory (Java VM heap) and freezes. In logging found:

WARN: [Caesium-1-2] [confluence.impl.hibernate.ConfluenceHibernateTransactionManager] doRollback Performing rollback. Transactions: ->[PluginReadWriteTX]: PROPAGATION-REQUIRED,ISOLATION_DEFAULT (session #561381725)

ERROR: [Caesium-1-2] [impl.schedule.caesium.JobRunnerWrapper] runJob Scheduled job com.atlassian.confluence.plugins.confluence-edge-index:flushEdgeIndexQueueJob#flushEdgeIndexQueue failed to run com.atlassian.bonnie.LuceneException: org.springframework.transaction.UnexpectedRollbackException: Transaction rolled back because it has been marked as rollback-only

Ilkka Kiiskinen

We have the same issue with multiple different instances with different users and add-ons. We have both docker and linux binary installations. Issues started when upgraded to 7.17.x versions.

We have an open issue to Atlassian support. Please open a support ticket to Atlassian to get more attention to this and to get support more data to find the root cause. Prepare your system to take java heap dumps.

We have already increased heap memory, but that seems not to be the solution.

Logs led us to this KB article: https://confluence.atlassian.com/confkb/confluence-logs-filled-with-the-error-failed-to-process-edge-index-task-838403476.html

We followed the instructions and rebuilt the content index from scratch including the optional step to restore popular content. https://confluence.atlassian.com/confkb/how-to-rebuild-the-content-indexes-from-scratch-on-confluence-server-110035351.html

Edge index errors stayed away only for awhile, so even this does not really fix the issue.

I hope we get some fix to this soon from Atlassian.