[
https://issues.apache.org/jira/browse/HIVE-13622?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eugene Koifman updated HIVE-13622:
----------------------------------
Description:
HIVE-13395 solves the the lost update problem with some inefficiencies.
1. TxhHandler.OperationType is currently derived from LockType. This doesn't
distinguish between Update and Delete but would be useful. See comments in
TxnHandler. Should be able to pass in Insert/Update/Delete info from client
into TxnHandler.
2. TxnHandler.addDynamicPartitions() should know the OperationType as well from
the client. It currently extrapolates it from TXN_COMPONENTS. This works but
requires extra SQL statements and is thus less performant. It will not work
multi-stmt txns. See comments in the code.
3. TxnHandler.checkLock() see more comments around
"isPartOfDynamicPartitionInsert". If TxnHandler knew whether it is being
called as part of an op running with dynamic partitions, it could be more
efficient. In that case we don't have to write to TXN_COMPONENTS at all during
lock acquisition. Conversely, if not running with DynPart then, we can kill
current txn on lock grant rather than wait until commit time.
All of these require some Thrift changes
Once done, re-enable TestDbTxnHandler2.testWriteSetTracking11()
was:
HIVE-13395 solves the the lost update problem with some inefficiencies.
1. TxhHandler.OperationType is currently derived from LockType. This doesn't
distinguish between Update and Delete but would be useful. See comments in
TxnHandler. Should be able to pass in Insert/Update/Delete info from client
into TxnHandler.
2. TxnHandler.addDynamicPartitions() should know the OperationType as well from
the client. It currently extrapolates it from TXN_COMPONENTS. This works but
requires extra SQL statements and is thus less performant. It will not work
multi-stmt txns. See comments in the code.
3. TxnHandler.checkLock() see more comments around
"isPartOfDynamicPartitionInsert". If TxnHandler knew whether it is being
called as part of an op running with dynamic partitions, it could be more
efficient. In that case we don't have to write to TXN_COMPONENTS at all during
lock acquisition. Conversely, if not running with DynPart then, we can kill
current txn on lock grant rather than wait until commit time.
All of these require some Thrift changes
> WriteSet tracking optimizations
> -------------------------------
>
> Key: HIVE-13622
> URL: https://issues.apache.org/jira/browse/HIVE-13622
> Project: Hive
> Issue Type: Bug
> Components: Transactions
> Affects Versions: 1.3.0, 2.1.0
> Reporter: Eugene Koifman
> Assignee: Eugene Koifman
>
> HIVE-13395 solves the the lost update problem with some inefficiencies.
> 1. TxhHandler.OperationType is currently derived from LockType. This doesn't
> distinguish between Update and Delete but would be useful. See comments in
> TxnHandler. Should be able to pass in Insert/Update/Delete info from client
> into TxnHandler.
> 2. TxnHandler.addDynamicPartitions() should know the OperationType as well
> from the client. It currently extrapolates it from TXN_COMPONENTS. This
> works but requires extra SQL statements and is thus less performant. It will
> not work multi-stmt txns. See comments in the code.
> 3. TxnHandler.checkLock() see more comments around
> "isPartOfDynamicPartitionInsert". If TxnHandler knew whether it is being
> called as part of an op running with dynamic partitions, it could be more
> efficient. In that case we don't have to write to TXN_COMPONENTS at all
> during lock acquisition. Conversely, if not running with DynPart then, we
> can kill current txn on lock grant rather than wait until commit time.
> All of these require some Thrift changes
> Once done, re-enable TestDbTxnHandler2.testWriteSetTracking11()
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
