[
https://issues.apache.org/jira/browse/HIVE-24291?focusedWorklogId=506248&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-506248
]
ASF GitHub Bot logged work on HIVE-24291:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 29/Oct/20 15:44
Start Date: 29/Oct/20 15:44
Worklog Time Spent: 10m
Work Description: klcopp commented on a change in pull request #1592:
URL: https://github.com/apache/hive/pull/1592#discussion_r514362309
##########
File path:
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/txn/CompactionTxnHandler.java
##########
@@ -281,9 +280,14 @@ public void markCompacted(CompactionInfo info) throws
MetaException {
try {
dbConn = getDbConn(Connection.TRANSACTION_READ_COMMITTED);
stmt = dbConn.createStatement();
- String s = "SELECT \"CQ_ID\", \"CQ_DATABASE\", \"CQ_TABLE\",
\"CQ_PARTITION\", " +
- "\"CQ_TYPE\", \"CQ_RUN_AS\", \"CQ_HIGHEST_WRITE_ID\" FROM
\"COMPACTION_QUEUE\" " +
- "WHERE \"CQ_STATE\" = '" + READY_FOR_CLEANING + "'";
+ /*
+ * By filtering on minOpenTxnWaterMark, we will only cleanup after
every transaction is committed, that could see
+ * the uncompacted deltas. This way the cleaner can clean up
everything that was made obsolete by this compaction.
+ */
+ long minOpenTxnWaterMark = getMinOpenTxnIdWaterMark(dbConn);
Review comment:
Cleaner already knows this value, Cleaner#run calls
CompactionTxnHandler#findMinOpenTxnIdForCleaner first, then findReadyToClean,
so you can just pass it into findReadyToClean.
(Btw findMinOpenTxnIdForCleaner doesn't filter out timed out txns like
getMinOpenTxnIdWaterMark does, might want to change that?
(AcidHouseKeeperService should take care of that, but who knows if it's on...
on the other hand that's another query and would take longer))
##########
File path:
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/txn/TxnHandler.java
##########
@@ -1526,6 +1529,10 @@ private void
updateWSCommitIdAndCleanUpMetadata(Statement stmt, long txnid, TxnT
if (txnType == TxnType.MATER_VIEW_REBUILD) {
queryBatch.add("DELETE FROM \"MATERIALIZATION_REBUILD_LOCKS\" WHERE
\"MRL_TXN_ID\" = " + txnid);
}
+ if (txnType == TxnType.COMPACTION) {
Review comment:
It's not the end of the world to add the CQ_TXN_ID column, but we can
avoid that and keep things more straightforward (i.e. keep compaction stuff out
of generic TxnHandler and limit it to CompactionTxnHandler which was made
specifically for updating compaction-related tables) by updating CQ_NEXT_TXN_ID
in CompactionTxnHandler instead, and calling it straight from Worker, maybe
right between commitTxn and markCompacted. It would be so much simpler.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 506248)
Time Spent: 40m (was: 0.5h)
> Compaction Cleaner prematurely cleans up deltas
> -----------------------------------------------
>
> Key: HIVE-24291
> URL: https://issues.apache.org/jira/browse/HIVE-24291
> Project: Hive
> Issue Type: Bug
> Reporter: Peter Varga
> Assignee: Peter Varga
> Priority: Major
> Labels: pull-request-available
> Time Spent: 40m
> Remaining Estimate: 0h
>
> Since HIVE-23107 the cleaner can clean up deltas that are still used by
> running queries.
> Example:
> * TxnId 1-5 writes to a partition, all commits
> * Compactor starts with txnId=6
> * Long running query starts with txnId=7, it sees txnId=6 as open in its
> snapshot
> * Compaction commits
> * Cleaner runs
> Previously min_history_level table would have prevented the Cleaner to delete
> the deltas1-5 until txnId=7 is open, but now they will be deleted and the
> long running query may fail if its tries to access the files.
> Solution could be to not run the cleaner until any txn is open that was
> opened before the compaction was committed (CQ_NEXT_TXN_ID)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
