[
https://issues.apache.org/jira/browse/HBASE-25478?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17266223#comment-17266223
]
Christine Feng commented on HBASE-25478:
----------------------------------------
[~stack] Here's some error logs we're seeing:
{{failed after 5 attempts,
“org.apache.hadoop.hbase.replication.ReplicationException: undeleted queue for
peerId: region_replica_replication”}}
It seems that when a test attempts to enable or create a table using this
"region_replica_replication" peerID, but the previous test using it has not
finished closing its tables yet, it causes a ReplicationException, which we
should catch, and retry the create/enable table operation until the previous
test successfully closes the table and frees up the peerID.
> Implement retries when enabling tables in TestRegionReplicaReplicationEndpoint
> ------------------------------------------------------------------------------
>
> Key: HBASE-25478
> URL: https://issues.apache.org/jira/browse/HBASE-25478
> Project: HBase
> Issue Type: Bug
> Reporter: Christine Feng
> Priority: Minor
>
> When a table from a previous test is too slow to close, it can cause problems
> for subsequent tests. Implementing a retry-and-wait function in this test
> class during table enablement will prevent issues where a single peerID is
> being used by two tests at once.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
