Details
-
Bug
-
Resolution: Fixed
-
Low
-
6.3.10
-
6.03
-
Description
Issues created on other nodes while one node is performing a foreground re-index will ultimately be lost from the index after the re-index operation is complete.
Steps to Reproduce
- Setup a JDC environment with at least 2 nodes
- Remove one of the nodes from the load balancer (so it is not accessible by normal users)
- Initiate a foreground re-index on the node you removed from the load balancer
- While the foreground re-index is in progress, create and update several issues on the other nodes(s) which are still in the load balancer and available to users
- Wait for the foreground re-index to complete, and for the new index to be completely replicated to all other nodes in the cluster
- Use the Issue Navigator (in List Mode, so it displays information from the index) to search for the issues you created and updated, and observe:
- Whether or not the issues appear in the search results
- What the last updated date/time is
- If the updates you made are reflected in the search results
Workaround
Run a full reindex during an outage window, and offline all of the other nodes while it runs, to prevent any possibility this will be encountered.
Attachments
Issue Links
- was cloned as
-
JDEV-32022 Loading...