Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Fix RestartIndexFollowingIT.testFollowIndex in case of fatal exception (
#92522) (#92649) This test failed several times after the leader cluster is fully restarted for the second time. The logs indicate that one or more ShardFollowNodeTask (the persistent task in charge or replication operations for a shard) have been stopped because a fatal exception occured. The fatal exception is an IllegalStateException with the Unable to open any connections to remote cluster message. I think this is due to the leader cluster being slow to restart and the remote cluster sniff strategy giving up after it tried to connect to the leader cluster nodes. Since this exception is fatal, the ShardFollowNodeTask stopped to replicate all operations and the test fails waiting for the number of docs to match on leader and follower clusters. The documented way to resolve CCR fatal exceptions for follower is to recreate the follower or to pause/resume follower. Test has been adjusted accordingly. Closes #90666 Co-authored-by: David Turner <david.turner@elastic.co> Co-authored-by: David Turner <david.turner@elastic.co>
- Loading branch information