Tungsten Replicator: data replication engine – MySQL master/slave server setup
————————-
switch back master/slave
————————-
Update: I forgot one crucial step when switching master/slave i.e. in cases of Master failover. That is to properly flush the log on Tungsten Replicator master .102 and note it’s log sequence number (118). And then on Tungsten Replicator slave which is to be promoted to master (in this case .101), tell it to wait until catch up to that sequence number (118).
[root@centos6 ~]# mysqladmin create test2 [root@centos6 ~]# mysql -e "show databases;" +--------------------+ | Database | +--------------------+ | information_schema | | mysql | | test | | test2 | | tungsten | | tungsten_rep1 | +--------------------+ [tungsten@centos6 tungsten]$ ./tungsten/tungsten-replicator/bin/trepctl services Processing services command... NAME VALUE ---- ----- appliedLastSeqno: 1 appliedLatency : 0.299 role : master serviceName : rep1 serviceType : local started : true state : ONLINE Finished services command... |
[tungsten@centos6 tungsten]$ ./tungsten/tungsten-replicator/bin/trepctl -service rep1 status Processing status command... NAME VALUE ---- ----- appliedLastEventId : mysql-bin.000006:0000000000000501;151 appliedLastSeqno : 1 appliedLatency : 0.299 clusterName : default currentEventId : mysql-bin.000006:0000000000000501 currentTimeMillis : 1316293854178 dataServerHost : 192.168.56.101 extensions : host : null latestEpochNumber : 0 masterConnectUri : thl://:/ masterListenUri : thl://192.168.56.101:2112/ maximumStoredSeqNo : 1 minimumStoredSeqNo : 0 offlineRequests : NONE pendingError : NONE pendingErrorCode : NONE pendingErrorEventId : NONE pendingErrorSeqno : -1 pendingExceptionMessage: NONE resourcePrecedence : 99 rmiPort : -1 role : master seqnoType : java.lang.Long serviceName : rep1 serviceType : local simpleServiceName : rep1 siteName : default sourceId : 192.168.56.101 state : ONLINE timeInStateSeconds : 2029.207 uptimeSeconds : 2034.258 Finished status command... |
* stop slave before master
[tungsten@localhost tungsten]$ mysql -e "show databases;" +--------------------+ | Database | +--------------------+ | information_schema | | mysql | | test | | test2 | | tungsten | | tungsten_rep1 | +--------------------+ [tungsten@localhost tungsten]$ ./tungsten/tungsten-replicator/bin/trepctl services Processing services command... NAME VALUE ---- ----- appliedLastSeqno: 1 appliedLatency : 26.72 role : slave serviceName : rep1 serviceType : local started : true state : ONLINE Finished services command... |
# Assign each server its new role
# 192.168.56.101 becomes the new master
[tungsten@localhost tungsten]$ ./tungsten/tungsten-replicator/bin/trepctl -service rep1 status Processing status command... NAME VALUE ---- ----- appliedLastEventId : mysql-bin.000006:0000000000000501;151 appliedLastSeqno : 1 appliedLatency : 26.72 clusterName : default currentEventId : NONE currentTimeMillis : 1316293891656 dataServerHost : 192.168.56.102 extensions : host : null latestEpochNumber : 0 masterConnectUri : thl://192.168.56.101:2112/ masterListenUri : thl://192.168.56.102:2112/ maximumStoredSeqNo : 1 minimumStoredSeqNo : 0 offlineRequests : NONE pendingError : NONE pendingErrorCode : NONE pendingErrorEventId : NONE pendingErrorSeqno : -1 pendingExceptionMessage: NONE resourcePrecedence : 99 rmiPort : -1 role : slave seqnoType : java.lang.Long serviceName : rep1 serviceType : local simpleServiceName : rep1 siteName : default sourceId : 192.168.56.102 state : ONLINE timeInStateSeconds : 2035.203 uptimeSeconds : 2038.652 Finished status command... |
# Assign 192.168.56.102 as new slave
# listening to 192.168.56.101 for data
[root@centos6 ~]# tail -10 /opt/tungsten/tungsten/tungsten-replicator/log/trepsvc.log INFO | jvm 1 | 2011/09/18 06:37:04 | 2011-09-18 06:37:04,791 [rep1 - Event dispatcher thread] INFO pipeline.Pipeline Starting pipeline: master INFO | jvm 1 | 2011/09/18 06:37:04 | 2011-09-18 06:37:04,792 [rep1 - binlog-to-q-0] INFO pipeline.SingleThreadStageTask Starting stage task thread INFO | jvm 1 | 2011/09/18 06:37:04 | 2011-09-18 06:37:04,792 [rep1 - Event dispatcher thread] INFO management.tungsten.TungstenPlugin Adding heartbeat for master INFO | jvm 1 | 2011/09/18 06:37:04 | 2011-09-18 06:37:04,796 [rep1 - q-to-thl-0] INFO pipeline.SingleThreadStageTask Starting stage task thread INFO | jvm 1 | 2011/09/18 06:37:04 | 2011-09-18 06:37:04,969 [rep1 - Event dispatcher thread] INFO management.OpenReplicatorManager Sent State Change Notification OFFLINE:NORMAL -> GOING-ONLINE:SYNCHRONIZING INFO | jvm 1 | 2011/09/18 06:37:04 | 2011-09-18 06:37:04,970 [ - WrapperSimpleAppMain] INFO management.ReplicationServiceManager internal/local replication service 'rep1' started successfully INFO | jvm 1 | 2011/09/18 06:37:04 | 2011-09-18 06:37:04,971 [rep1 - Event dispatcher thread] INFO management.OpenReplicatorManager Sent State Change Notification GOING-ONLINE:SYNCHRONIZING -> ONLINE INFO | jvm 1 | 2011/09/18 06:37:13 | 2011-09-18 06:37:13,026 [rep1 - ConnectorHandler: initializing] INFO thl.Protocol THL protocol buffering enabled: size=10 INFO | jvm 1 | 2011/09/18 06:37:13 | 2011-09-18 06:37:13,034 [rep1 - ConnectorHandler: initializing] INFO thl.ConnectorHandler New THL client connection: sourceID=192.168.56.102 heartbeatMillis=3000 INFO | jvm 1 | 2011/09/18 06:37:13 | 2011-09-18 06:37:13,034 [rep1 - ConnectorHandler: initializing] INFO thl.ConnectorHandler Client log checking disabled; not checking for diverging histories |
# start both servers
trepctl -service rep1 -host 192.168.56.101 online trepctl -service rep1 -host 192.168.56.102 online |
# check service for new roles
trepctl -host 192.168.56.101 services trepctl -host 192.168.56.102 services |
[tungsten@centos6 tungsten]$ trepctl -host 192.168.56.101 services Processing services command... NAME VALUE ---- ----- appliedLastSeqno: 14 appliedLatency : 0.39 role : master serviceName : rep1 serviceType : local started : true state : ONLINE Finished services command... |
[tungsten@localhost root]$ trepctl -host 192.168.56.102 services Processing services command... NAME VALUE ---- ----- appliedLastSeqno: 14 appliedLatency : 12.393 role : slave serviceName : rep1 serviceType : local started : true state : ONLINE Finished services command... |
This is as far as I have progressed and will play with Tungsten Replicator some more. I’d like to thank Giuseppe Maxia for his wonderful Tungsten Replicator Recipe Cookbook in helping me get started.
Tungsten Replicator can also replicate to other SQL databases, MySQL replication to PostgreSQL, Oracle, MongoDB etc. Zappo.com online retailer posted a Youtube video on how they use Tungsten Replicator to replicate data between MySQL and Oracle.
Other useful Youtube videos where Giuseppe Maxia explains Tungsten Replicator’s Global Transaction IDs and Advanced replication for the Masses.
