Add some labels to evaluate what are the exactly cost of tx.concurrencyMgr().lockRecordFileHeader, and found out that most of the time are basically come from xLocks waiting while loop. However, elasql don't actually use this LockTable, so the change in this table is actually of no use.
Add some labels to evaluate what are the exactly cost of tx.concurrencyMgr().lockRecordFileHeader, and found out that most of the time are basically come from xLocks waiting while loop. However, elasql don't actually use this LockTable, so the change in this table is actually of no use.