Hey , i notice some deadlocks trigered by PoolService and PoolServer.
This happen when PoolServer._flushSharesToDb() run in same time with
PoolService._removeOldDbEntries() .
LATEST DETECTED DEADLOCK
------------------------
<>
*** (1) TRANSACTION:
TRANSACTION <>, ACTIVE 0 sec fetching rows
mysql tables in use 2, locked 2
LOCK WAIT 5 lock struct(s), heap size 1184, 237 row lock(s), undo log entries 12
MySQL thread id <>, OS thread handle <>, query id <> 127.0.0.1 <SERVICE> Sending data
DELETE shares
FROM shares
LEFT JOIN block on shares.prev_block=block.id
WHERE block.height < <>
*** (1) WAITING FOR THIS LOCK TO BE GRANTED:
RECORD LOCKS space id 3127 page no 3 n bits 400 index `PRIMARY` of table `pool`.`shares` trx table locks 2 total table locks 2 trx id <>lock_mode X waiting lock hold time 0 wait time before grant 0
*** (2) TRANSACTION:
TRANSACTION <>, ACTIVE 0 sec starting index read, thread declared inside InnoDB 4993
mysql tables in use 1, locked 1
4 lock struct(s), heap size 1184, 6 row lock(s), undo log entries 2
MySQL thread id 9, OS thread handle <>, query id <>127.0.0.1 <SERVER>update
INSERT INTO shares (user, device, prev_block, count, difficulty)
VALUES (<>), (<>), (<>), (<>), (<>), (<>), (<>) ON DUPLICATE KEY UPDATE count=count+values(count), difficulty=difficulty+values(difficulty)
*** (2) HOLDS THE LOCK(S):
RECORD LOCKS space id 3127 page no 3 n bits 400 index `PRIMARY` of table `pool`.`shares` trx table locks 1 total table locks 2 trx id <> lock_mode X locks rec but not gap lock hold time 0 wait time before grant 0
*** (2) WAITING FOR THIS LOCK TO BE GRANTED:
RECORD LOCKS space id 3127 page no 3 n bits 400 index `PRIMARY` of table `pool`.`shares` trx table locks 1 total table locks 2 trx id <>lock_mode X locks rec but not gap waiting lock hold time 0 wait time before grant 0
*** WE ROLL BACK TRANSACTION (2)
Hey , i notice some deadlocks trigered by PoolService and PoolServer.
This happen when
PoolServer._flushSharesToDb()run in same time withPoolService._removeOldDbEntries().Mysql Details:
Are conflict transactions re executed ? Or missing value data there ?
Any way to avoid deadlocks or at least workaround when this happening ?
P.S i removed some sensitive data where you see
<>