pvargacl commented on a change in pull request #1839:
URL: https://github.com/apache/hive/pull/1839#discussion_r553598695
##########
File path:
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/txn/CompactionTxnHandler.java
##########
@@ -993,7 +994,7 @@ public boolean checkFailedCompactions(CompactionInfo ci)
throws MetaException {
try {
try {
dbConn = getDbConn(Connection.TRANSACTION_READ_COMMITTED);
- pStmt = dbConn.prepareStatement("SELECT \"CC_STATE\" FROM
\"COMPLETED_COMPACTIONS\" WHERE " +
+ pStmt = dbConn.prepareStatement("SELECT \"CC_STATE\",
\"CC_ENQUEUE_TIME\" FROM \"COMPLETED_COMPACTIONS\" WHERE " +
Review comment:
enqueue time is guaranteed to be filled out for every compaction, the
others might be null for a failed one.
##########
File path:
standalone-metastore/metastore-common/src/main/java/org/apache/hadoop/hive/metastore/conf/MetastoreConf.java
##########
@@ -415,6 +415,10 @@ public static ConfVars getMetaConf(String name) {
new RangeValidator(1, 20), "Number of consecutive compaction failures
(per table/partition) " +
"after which automatic compactions will not be scheduled any more.
Note that this must be less " +
"than hive.compactor.history.retention.failed."),
+
COMPACTOR_INITIATOR_FAILED_RETRY_TIME("metastore.compactor.initiator.failed.retry.time",
+ "hive.compactor.initiator.failed.retry.time", 7, TimeUnit.DAYS,
+ "Time after Initiator will ignore
metastore.compactor.initiator.failed.compacts.threshold and retry with
compaction again. \n"
Review comment:
fixed
##########
File path:
standalone-metastore/metastore-server/src/main/java/org/apache/hadoop/hive/metastore/txn/CompactionTxnHandler.java
##########
@@ -1006,16 +1007,23 @@ public boolean checkFailedCompactions(CompactionInfo
ci) throws MetaException {
rs = pStmt.executeQuery();
int numFailed = 0;
int numTotal = 0;
+ long lastEnqueueTime = -1;
int failedThreshold = MetastoreConf.getIntVar(conf,
ConfVars.COMPACTOR_INITIATOR_FAILED_THRESHOLD);
while(rs.next() && ++numTotal <= failedThreshold) {
+ long enqueueTime = rs.getLong(2);
+ if (enqueueTime > lastEnqueueTime) {
+ lastEnqueueTime = enqueueTime;
+ }
if(rs.getString(1).charAt(0) == FAILED_STATE) {
numFailed++;
}
else {
numFailed--;
}
}
- return numFailed == failedThreshold;
+ // If the last attempt was too long ago, ignore the failed treshold
and try compaction again
Review comment:
fixed
----------------------------------------------------------------
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:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
