The problem arises because the bulk transition code updates the Issue object, then runs the validator, and irrespective of the validator results, reindexes the issue. So although the post-functions don't get run (because a validator failed) the modified Issue index is reindex, persisting any field updates (eg. resolution) to the index.

The solution is to not reindex issues in the bulk transition code, but rather rely on the reindexing post-function.

I've tested this with 3.11 and 3.12 and both returned the same error:

2008-02-13 12:26:55,591 http-8090-Processor20 ERROR [atlassian.jira.workflow.SimpleWorkflowManager] An exception occurred
[InvalidInputException: [Error map: [{}]] [Error list: [[User 'admin' doesn't have the 'worklogdeleteall' permission]]]
        at com.atlassian.jira.workflow.validator.AbstractPermissionValidator.hasUserPermission(AbstractPermissionValidator.java:42)
        at com.atlassian.jira.workflow.validator.PermissionValidator.validate(PermissionValidator.java:67)

However, my results were different from the description. After transitioning the issue via bulk and breaking the validation, the issue did not move on to the next status and the index was not updated. A re-index did not do anything.

It does not display a proper error message to the user that the process failed.

The errors above appeared in the log file, right? They did not appear in the UI.

It does not display a proper error message to the user that the process failed.

Yes, this is a bug tracked by JRA-9844.

Therefore, as the index is not corrupted it looks like this issue is no longer a problem. This issue is only about corrupting the index when validators fail.

Cheers,
Anton

Can you let me know if you agree?

Cheers,
Anton

Yes, this would be the same as JRA-9844. The index is not corrupted.

Also this error still happens on version 3.11 and 3.12 which could be an addition for JRA-9844.

Regards,
Timothy

I will resolve this issue.

Cheers,
Anton