Bug #14396 | replication: insert data from one db to another db not existing on the slave | ||
---|---|---|---|
Submitted: | 27 Oct 2005 14:22 | Modified: | 29 Oct 2005 7:43 |
Reporter: | Armin Lorenz | Email Updates: | |
Status: | Not a Bug | Impact on me: | |
Category: | MySQL Server | Severity: | S2 (Serious) |
Version: | 4.1.15 5.0.15 | OS: | W2000 Server |
Assigned to: | Guilhem Bichot | CPU Architecture: | Any |
[27 Oct 2005 14:22]
Armin Lorenz
[27 Oct 2005 17:24]
Jonathan Miller
Description: Master 4.1.7 Slave 4.1.15 und 5.0.15 Why is the master not updated to at least 4.1.15?
[28 Oct 2005 5:04]
Armin Lorenz
Its a industrial production 24/7 solution. Did not find any fixed error in the following important to me. And so never touch a running system. Please correct me if i oversee the problem described in this report. Slave was easier to upgrade.
[28 Oct 2005 11:46]
Jonathan Miller
For "a industrial production 24/7 solution." it usually best to keep all version at the same rev level. Thanks JBM
[28 Oct 2005 11:53]
Armin Lorenz
Ok. This is your opinion. But does this belong to the bug?
[28 Oct 2005 12:00]
Armin Lorenz
I dont want to be rough. But there are several reasons for not upgrading the master without need. The first time the problem showed up both servers are on the same old version. I tried out the 4.1.15 on the slave because i can shutdown this server because its only load balance. But the problem did remain. I cannot shutdown the master so easily.
[29 Oct 2005 7:43]
Guilhem Bichot
Hello Armin, What you experience is the expected and documented behaviour of replicate-do-db (so I'm closing this report as "not a bug"). Slave has "replicate-do-db=ipc_tlm" so any master's binlog event which has ipc_tlm as its *default* database will be executed. This is the case of this event: 051027 16:00:57 [ERROR] Slave: Error 'Unknown database 'ipc_tlm_backup'' on query. Default database: 'ipc_tlm'. Query: 'create table ipc_tlm_backup.notused_tblstatistic3 select * from ipc_tlm.tblstatistic where sreason = 7 and toolsid < 0', Error_code: 1049 It's as documented here: http://dev.mysql.com/doc/refman/4.1/en/binary-log.html Two solutions: 1) do a "USE ipc_tlm_backup" on your master's connection before doing the CREATE TABLE, this way the default db will be ipc_tlm_backup and so the event will be ignored by the slave or 2) don't use replicate-do-db and use replicate-wild-do-table=ipc_tlm.%