Luke Chen created KAFKA-13033:
---------------------------------
Summary: coordinator not available error should cause add into
unmap list to do a new lookup
Key: KAFKA-13033
URL: https://issues.apache.org/jira/browse/KAFKA-13033
Project: Kafka
Issue Type: Bug
Affects Versions: 3.0.0
Reporter: Luke Chen
Assignee: Luke Chen
In KIP-699, we add some handler to handle different types of operation. In the
`handleError`, we didn't make the `COORDINATOR_NOT_AVAILABLE` as unmapped, to
do a re-lookup. In
[DescribeTransactionsHandler|[https://github.com/apache/kafka/blob/trunk/clients/src/main/java/org/apache/kafka/clients/admin/internals/DescribeTransactionsHandler.java#L172-L186],]
[~hachikuji] has explained why `COORDINATOR_NOT_AVAILABLE` and
`NOT_COORDINATOR` should be listed in unmapped, and
`COORDINATOR_LOAD_IN_PROGRESS` should not.
{code:java}
case COORDINATOR_LOAD_IN_PROGRESS:
// If the coordinator is in the middle of loading, then we just need to
retry
log.debug("DescribeTransactions request for transactionalId `{}` failed
because the " +
"coordinator is still in the process of loading state. Will retry",
transactionalIdKey.idValue);
break;
case NOT_COORDINATOR:
case COORDINATOR_NOT_AVAILABLE:
// If the coordinator is unavailable or there was a coordinator change,
then we unmap
// the key so that we retry the `FindCoordinator` request
unmapped.add(transactionalIdKey);
log.debug("DescribeTransactions request for transactionalId `{}` returned
error {}. Will attempt " +
"to find the coordinator again and retry",
transactionalIdKey.idValue, error);
break;
{code}
We should be consistent with it. Fix it, and the tests.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
