MySQL 5.6.21
Galera 25.3.5
Percona-xtrabackup-21 2.1.9-746-1.precise
Got error: 1047
Fix:
Database is probably in ‘initialized’ state. need to restart the service and check configuration.
WSREP_SST: [ERROR] Error while getting data from donor node: exit codes: 1 0 (20150325 09:17:28.755) WSREP_SST: [ERROR] Cleanup after exit with status:32 (20150325 09:17:28.756) WSREP_SST: [INFO] Removing the sst_in_progress file (20150325 09:17:28.759) 2015-03-25 09:17:28 6459 [ERROR] WSREP: Process completed with error: wsrep_sst_xtrabackup --role 'joiner' --address 'ip_address' --auth 'user:password' --datadir '/var/lib/mysql/' --defaults-file '/etc/mysql/my.cnf' --defaults-group-suffix '' --parent '6459' --binlog 'binlog' : 32 (Broken pipe) 2015-03-25 09:17:28 6459 [ERROR] WSREP: Failed to read uuid:seqno from joiner script. 2015-03-25 09:17:28 6459 [ERROR] WSREP: SST failed: 32 (Broken pipe) 2015-03-25 09:17:28 6459 [ERROR] Aborting 2015-03-25 09:17:28 6459 [Warning] WSREP:State transfer failed: -2 (No such file or directory) 2015-03-25 09:17:28 6459 [ERROR] WSREP: gcs/src/gcs_group.c:gcs_group_handle_join_msg():723: Will never receive state. Need to abort.
Fix:
Check innodb_data_file_path=ibdata1:256M:autoextend does not exist in my.cnf
Check that xtrabackup isn’t still running orphaned on either node “ps aux | grep mysql” on donor
Check apt-get install persona-xtrabackup-21 is installed (All servers should be running the same version)
Check ssh keys are correct and servers can ssh freely between each other
Make sure sst_in_progress file does not exist on the joining server
Check that the node joining has an /etc/hosts file entry for the donor (or DNS is good)
Check the wsrep_sst_method on the donor at runtime is xtrabackup-v2
Make sure the root:password in /etc/mysql/my.cnf is the same as what the localhost is or xtrabackup can’t use it and will bomb
[ERROR] Unknown/unsupported storage engine: InnoDB 2015-03-25 14:33:48 31707 [ERROR] Aborting
Fix:
ibdata and log files are probably not the same size as the ones on the donor server. If innodb_data_file_path is set, this could be causing the problem.
Other interesting facts:
Syncing with xtrabackup:
– creating or dropping a database during sync will cause the node syncing to the cluster to drop