We have a large database(40tb) and buffer pool (128gb), with heavy read/write. During peak hours, we get lockups every 10 minutes or so. The lockup lasts about 10 seconds, and during the lockup all disk activity drops to 0. The mysql processlist shoots up extremely high and all queries look like they are locked. After 10 seconds or so, everything goes back to normal.
We are running version 5.1.45-51 with Percona SQL Server (GPL), XtraDB 10.2
The interesting this is during the lock period, show innodb status gives lots of semaphores like this:
Thread 1251916096 has waited at btr/btr0cur.c line 487 for 28.000 seconds the semaphore:
S-lock on RW-latch at 0×2ac843e8e770 '&new_index->lock'
a writer (thread id 1265760576) has reserved it in mode exclusive
number of readers 0, waiters flag 1, lock_word: 0
Last time read locked in file btr/btr0cur.c line 487
Last time write locked in file btr/btr0cur.c line 480
From what i've read I think it's buffer pool contention. Has anyone seen this before?
We are running version 5.1.45-51 with Percona SQL Server (GPL), XtraDB 10.2
The interesting this is during the lock period, show innodb status gives lots of semaphores like this:
Thread 1251916096 has waited at btr/btr0cur.c line 487 for 28.000 seconds the semaphore:
S-lock on RW-latch at 0×2ac843e8e770 '&new_index->lock'
a writer (thread id 1265760576) has reserved it in mode exclusive
number of readers 0, waiters flag 1, lock_word: 0
Last time read locked in file btr/btr0cur.c line 487
Last time write locked in file btr/btr0cur.c line 480
From what i've read I think it's buffer pool contention. Has anyone seen this before?
< AF35FED2 undo n
Comment