Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Incomplete
-
10.0.10
-
None
-
None
-
CentOS x86_64 (master) and CentOS i386 (slave)
Description
I've used MariaDB 5.5 for month with replication without a hiccup. As the upgrade plan sees, I've upgraded the slave to 10 and then the master to 10 also and from that moment I'm seeing constantly increasing slave lag...
From iotop shows
TID PRIO USER DISK READ DISK WRITE SWAPIN IO> COMMAND
|
337 be/3 root 0.00 B/s 3.89 K/s 0.00 % 89.96 % [jbd2/md1-8]
|
3313 be/4 mysql 11.66 K/s 116.61 K/s 0.00 % 5.59 % mysqld --~-port=3306
|
and jbd2 clogs completely IO... witch didn't happen... 
Checking process list, and refreshing it constantly, I see constantly Table lock in COMMIT phase...
+----+-----------------+-----------------+---------+---------+--------+----------------------------------+------------------+----------+
|
| Id | User | Host | db | Command | Time | State | Info | Progress |
|
+----+-----------------+-----------------+---------+---------+--------+----------------------------------+------------------+----------+
|
| 1 | event_scheduler | localhost | NULL | Daemon | 3255 | Waiting on empty queue | NULL | 0.000 |
|
| 10 | system user | | NULL | Connect | 1991 | Waiting for master to send event | NULL | 0.000 |
|
| 11 | system user | | MCPC_SI | Connect | 127178 | Table lock | COMMIT | 0.000 |
|
| 15 | root | localhost | NULL | Sleep | 0 | | NULL | 0.000 |
|
| 16 | root | localhost:41879 | MCPC_SI | Sleep | 15 | | NULL | 0.000 |
|
| 17 | root | localhost | NULL | Query | 0 | init | SHOW PROCESSLIST | 0.000 |
|
+----+-----------------+-----------------+---------+---------+--------+----------------------------------+------------------+----------+
|
I've changed from innodb to myisam (it could be a storage engine bottleneck) but without any changes...
Any clues on what might be (besides the change from 5.5 to 10)?
Thank you,
António Fernandes