[
https://issues.apache.org/jira/browse/HIVE-25943?focusedWorklogId=737413&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-737413
]
ASF GitHub Bot logged work on HIVE-25943:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 07/Mar/22 08:14
Start Date: 07/Mar/22 08:14
Worklog Time Spent: 10m
Work Description: veghlaci05 commented on a change in pull request #3034:
URL: https://github.com/apache/hive/pull/3034#discussion_r820464621
##########
File path:
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/txn/CompactionTxnHandler.java
##########
@@ -353,11 +353,9 @@ public void markCompacted(CompactionInfo info) throws
MetaException {
if (minOpenTxnWaterMark > 0) {
whereClause += " AND (\"CQ_NEXT_TXN_ID\" <= " + minOpenTxnWaterMark
+ " OR \"CQ_NEXT_TXN_ID\" IS NULL)";
}
- if (retentionTime > 0) {
- whereClause += " AND \"CQ_COMMIT_TIME\" < (" + getEpochFn(dbProduct)
+ " - " + retentionTime + ")";
- }
+ whereClause += " AND (\"CQ_COMMIT_TIME\" < (" + getEpochFn(dbProduct)
+ " - CQ_RETRY_RETENTION - " + retentionTime + ") OR \"CQ_COMMIT_TIME\" IS
NULL)";
Review comment:
This is my plan, but unfortunately this change cannot be separated.
>From now on the CQ_COMMIT_TIME comparisaon needs to be done all the case
(because of CQ_RETRY_RETENTION), regardless of the value of the retention time.
If I would keep the if condition, retry retention would not work if the cleaner
retention time is 0. What I plan to do is to fix the where clause in a separate
ticket too, and this PR should be merged to master only after that one,
mimicking that the fix (adding the _OR "CQ_COMMIT_TIME" IS NULL_ part) existed
before, and this PR built on top of that fix. What do you think?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: gitbox-unsubscr...@hive.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 737413)
Time Spent: 5h 10m (was: 5h)
> Introduce compaction cleaner failed attempts threshold
> ------------------------------------------------------
>
> Key: HIVE-25943
> URL: https://issues.apache.org/jira/browse/HIVE-25943
> Project: Hive
> Issue Type: Improvement
> Components: Hive
> Reporter: László Végh
> Assignee: László Végh
> Priority: Major
> Labels: pull-request-available
> Time Spent: 5h 10m
> Remaining Estimate: 0h
>
> If the cleaner fails for some reason, the compaction entity status remains in
> "ready for cleaning", therefore the cleaner will pick up this entity
> resulting in an endless try. The number of failed cleaning attempts should be
> counted and if they reach a certain threshold the cleaner must skip all the
> cleaning attempts on that compaction entity.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
