[
https://issues.apache.org/jira/browse/HIVE-26569?focusedWorklogId=831488&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-831488
]
ASF GitHub Bot logged work on HIVE-26569:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 06/Dec/22 16:50
Start Date: 06/Dec/22 16:50
Worklog Time Spent: 10m
Work Description: abstractdog commented on code in PR #3626:
URL: https://github.com/apache/hive/pull/3626#discussion_r1041219149
##########
llap-tez/src/java/org/apache/hadoop/hive/llap/tezplugins/LlapTaskCommunicator.java:
##########
@@ -584,6 +581,17 @@ public void indicateError(Throwable t) {
});
}
+ /**
+ * This is a best effort action in case of an invalid LLAP_TOKEN. The
communicator (LlapProtocolClientProxy) periodically
+ * refreshes tokens, but there is a relatively small time window (compared
to the whole token maxLifetime),
+ * when the token is cancelled and it's not yet fetched from daemons. By the
time we detect this problem, the current
+ * task attempt is already failed, but usually there are at least 3 task
attempts before failing the task and dag, therefore
+ * fetching an LLAP_TOKEN synchronously here definitely solves the invalid
token problem (as the next task attempt will use it).
+ */
+ private void handleInvalidToken(Throwable t) {
+ this.communicator.refreshToken();
Review Comment:
ack, it's just a leftover
Issue Time Tracking
-------------------
Worklog Id: (was: 831488)
Time Spent: 2.5h (was: 2h 20m)
> Support renewal and recreation of LLAP_TOKENs
> ---------------------------------------------
>
> Key: HIVE-26569
> URL: https://issues.apache.org/jira/browse/HIVE-26569
> Project: Hive
> Issue Type: Improvement
> Reporter: László Bodor
> Assignee: László Bodor
> Priority: Major
> Labels: pull-request-available
> Time Spent: 2.5h
> Remaining Estimate: 0h
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
