smengcl commented on code in PR #4280:
URL: https://github.com/apache/ozone/pull/4280#discussion_r1129806256
##########
hadoop-ozone/ozone-manager/src/main/java/org/apache/hadoop/ozone/om/OmMetadataManagerImpl.java:
##########
@@ -549,6 +564,8 @@ protected void initializeOmTables(boolean addCacheMetrics)
deletedTable = this.store.getTable(DELETED_TABLE, String.class,
RepeatedOmKeyInfo.class);
checkTableStatus(deletedTable, DELETED_TABLE, addCacheMetrics);
+ // Currently, deletedTable is the only table that will need the table lock
Review Comment:
It looks to me the parameter for the `deletedTable` put operation in
`OMDirectoriesPurgeResponseWithFSO` could be wrong. It does not seem to be
considering the case where `deletedKey` could have already existed in the
`deletedTable`:
https://github.com/apache/ozone/blob/5eead92666b374c4bb912f332488acb23f9a81df/hadoop-ozone/ozone-manager/src/main/java/org/apache/hadoop/ozone/om/response/key/OMDirectoriesPurgeResponseWithFSO.java#L114-L122
The correct action would be to append the key to the existing
`RepeatedOmKeyInfo` value in the `deletedTable` in this case.
This reminds me of HDDS-5905 which is recently brought up again.
Or it could be that because it is FSO this would never happen because of the
Object ID prefix.
Regarding the lock, I will put a TODO here at the moment. We will need to
revisit the locking part in HDDS-8067.
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
