- Master version: mysql-server-5.5.14-1.el5.remi
- Slave version: Percona-XtraDB-Cluster-server-5.5.24-23.6.340
- Binlog format: ROW based
I'm using PRM to setup HA for MySQL.
Sometime, MySQL slave stopped with some errors:
Error executing row event: 'Table 'reportingdb.tvc_ads_tag_date' doesn't exist'
[Warning] Slave SQL: Could not execute Update_rows event on table reportingdb.7k_banner_channel_tmp; Can't find record in '7k_banner_channel_tmp', Error_code: 1032; handler error HA_ERR_KEY_NOT_FOUND; the event's master log mysql-bin.000854, end_log_pos 859094925, Error_code: 1032
120828 0:36:13 [Warning] Slave SQL: Could not execute Write_rows event on table reportingdb.7k_bookings_ver; Duplicate entry '1518' for key 'PRIMARY', Error_code: 1062; handler error HA_ERR_FOUND_DUPP_KEY; the event's master log mysql-bin.001022, end_log_pos 509479, Error_code: 1062
Error 'Error in list of partitions to DROP' on query. Default database: 'reportingdb'. Query: 'ALTER TABLE v3_ban_date_cpm7k DROP PARTITION pcurrent_201298'
mysql.sock is missing but the MySQL processes still exist. Pacemaker cannot restart due to the below errors:
InnoDB: Unable to lock ./ibdata1, error: 11 InnoDB: Check that you do not already have another mysqld process InnoDB: using the same InnoDB data or log files.
I have to login via TCP (
-h 127.0.0.1) and stop MySQL first, then let
corosync start MySQL again.
In summary, the question is:
- How do I solve the above slave errors and prevent them from happening in the future?
- Why is the
mysql.sockmissing? How do I make the Pacemaker better handle this situation?