0

After a Schema-Change on which I have missed out to take care about replication, I now have one Server not synching anymore, as it tries to alter the table which already has been taken care of (Server 1&2 have the additional column and are fine, Server 3 tries to insert it again - errors and therewith is in wsrep_local_state: 2

The error log is quite clear with that:

150728 15:19:29 [ERROR] Slave SQL: Error 'Duplicate column name 'somecolumn'' on query. Default database: 'my_db'. Query: 'ALTER TABLE tablename ADD somecolumn DATETIME NULL', Internal MariaDB error code: 1060
150728 15:19:29 [Warning] WSREP: RBR event 1 Query apply warning: 1, 195651391
150728 15:19:29 [Warning] WSREP: Ignoring error for TO isolated action: source: f7715036-20ca-11e5-a90f-f23b6463dd16 version: 3 local: 0 state: APPLYING flags: 65 conn_id: 2348812 trx_id: -1 seqnos (l: 14350890, g: 195651391, s: 195651390, d: 195651390, ts: 20916266579392088)

As it is quite obvious this transaction can be skipped, I would like to do so - but can't really find anywhere how to (only for a master-slave setup with a "SQL_SLAVE_SKIP_COUNTER=1" - which won't work on galera multi-master).

Chris
  • 123
  • 1
  • 4

1 Answers1

0

The issue is - generally - resolved. I restarted the other servers to make sure there are no error- messages, then restarted the Server in question (the one not synching) - again. The "out-of-sync" stated must have happened on re- synchronisation of the servers (I tried twice...), and possibly there was still something pending on one of the other servers.

Chris
  • 123
  • 1
  • 4