[jira] [Commented] (HIVE-22755) Cleaner/Compaction can skip the read locks and use the min open txn id
[ https://issues.apache.org/jira/browse/HIVE-22755?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17020942#comment-17020942 ] Peter Vary commented on HIVE-22755: --- In my deleted comment I misunderstood the purpose of the jira :( The read lock is needed so there is no concurrent DDL command running parallel the compaction. Before having versioned table/partition metadata, removing this read lock could be problematic. > Cleaner/Compaction can skip the read locks and use the min open txn id > -- > > Key: HIVE-22755 > URL: https://issues.apache.org/jira/browse/HIVE-22755 > Project: Hive > Issue Type: Sub-task > Components: Transactions >Reporter: Slim Bouguerra >Priority: Major > Fix For: 4.0.0 > > > The minOpenTxnId is used by the Cleaner here > https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/txn/compactor/Cleaner.java#L154 > This currently converts it to open write-ids to clean appropriately. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-22755) Cleaner/Compaction can skip the read locks and use the min open txn id
[ https://issues.apache.org/jira/browse/HIVE-22755?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17020832#comment-17020832 ] Peter Vary commented on HIVE-22755: --- I have 2 ideas: * Time based approach: use the start time of the minOpenTxn to filter the files to clean * Instead of skipping every read lock, create a read lock per table instead of per partition. This would fix some failures created by the previous patch with the added benefit of knowing what tables are under read. CC: [~lpinter], [~dkuzmenko] > Cleaner/Compaction can skip the read locks and use the min open txn id > -- > > Key: HIVE-22755 > URL: https://issues.apache.org/jira/browse/HIVE-22755 > Project: Hive > Issue Type: Sub-task > Components: Transactions >Reporter: Slim Bouguerra >Priority: Major > Fix For: 4.0.0 > > > The minOpenTxnId is used by the Cleaner here > https://github.com/apache/hive/blob/master/ql/src/java/org/apache/hadoop/hive/ql/txn/compactor/Cleaner.java#L154 > This currently converts it to open write-ids to clean appropriately. -- This message was sent by Atlassian Jira (v8.3.4#803005)
[jira] [Commented] (HIVE-22755) Cleaner/Compaction can skip the read locks and use the min open txn id
[ https://issues.apache.org/jira/browse/HIVE-22755?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=17020594#comment-17020594 ] Slim Bouguerra commented on HIVE-22755: --- cc [~t3rmin4t0r] please feel free to add more insights about your idea on how the Cleaner can skip the read lock. > Cleaner/Compaction can skip the read locks and use the min open txn id > -- > > Key: HIVE-22755 > URL: https://issues.apache.org/jira/browse/HIVE-22755 > Project: Hive > Issue Type: Sub-task >Reporter: Slim Bouguerra >Priority: Major > -- This message was sent by Atlassian Jira (v8.3.4#803005)