2 clusters - one master, one slave - problem

[Grzegorz Greg]

Hello.

I've setup a cluster with 2 mgmd, 6 ndb and 6 sql nodes. All working brilliant. I wanted to create backup cluster in other geographical localization. Not a circ replication or multi master. Just master -> slave. So I've created the same second cluster (2 mgmd, 6 ndb and 6 sql nodes), on master cluster on each sql node I turn on binlog, setup bin-do-db. I've backup master cluster, then I restore backups on slave cluster. I started slaves and replication working but every time on each slave sql node I get errors:

121015 15:14:41 [ERROR] Slave SQL: Could not execute Write_rows event on table files; Got temporary error 266 'Time-out in NDB, probably caused by
deadlock' from NDB, Error_code: 1297; Lock wait timeout exceeded; try restarting transaction,
Error_code: 1205; handler error HA_ERR_LOCK_WAIT_TIMEOUT; the event's master log mysql-bin.000156,
end_log_pos 66749389, Error_code: 1205


The similar errors exists on each sql nodes on slave cluster.


Please tell me what can be wrong? Does each sql master node have to be a master for each sql slave node? Now it is. Also each master have own server-id and each slave have own server-id. There's no NDB errors on master cluster, only on slave.

Here's a part of my.cnf on each master sql node:

log_bin = mysql-bin
binlog-format = row
expire_logs_days = 1
max_binlog_size = 100M
server-id = 1
binlog-do-db=db1
binlog-do-db=db2


Here's a part of my.cnf on each slave sql node:


#log_bin = mysql-bin
#binlog-format = row
expire_logs_days = 1
#max_binlog_size = 100M
server-id = 11
replicate-do-db=db1
replicate-do-db=db2


Thank you for your help.

Best regards.

Greg