Operation examples for shared wal archives environment - t-matsuo/resource-agents GitHub Wiki
(in the writing)
I'm not recommend sharing wal-archives because it's difficult to keep consisntency of wal-archives between pm01 and pm02. Inconsistent wal-archives may break data.
I hope someone will make great script to be set in archive_command.
- start pacemaker on pm01 and pm02
- pm01 and pm02 become Slave
- If /var/lib/pgsql/tmp/PGSQL.lock is exist, it means that data is inconsistent. So RA don't start PostgreSQL.
- RA begins to compare xlog-location between pm01 and pm02
- assume pm01's data is newer than pm02's
- wait to be Master on pm01
- on pm01, pgsql-status becomes "PRI" and pgsql-data-status becomes "LATEST"
- on pm02, pgsql-status becomes "HS:sync" and pgsql-data-status becomes "STREAMING|SYNC"
- If pgsql-status keeps "HS:alone" and pgsql-data-status keeps "DISCONNECTED" on pm02, it means that PostgreSQL cannot be replication mode. Please see PostgreSQL logs.
- pm01 and pm02 become Slave
(display example)
# crm_mon -Af
============
Last updated: Wed Feb 15 15:15:15 2012
Stack: Heartbeat
Current DC: pm02 (22222222-2222-2222-2222-222222222222) - partition with quorum
Version: 1.0.11-1554a83db0d3c3e546cfd3aaff6af1184f79ee87
2 Nodes configured, unknown expected votes
3 Resources configured.
============
Online: [ pm01 pm02 ]
Resource Group: master-group
vip-master (ocf::heartbeat:IPaddr2): Started pm01
vip-rep (ocf::heartbeat:IPaddr2): Started pm01
Master/Slave Set: msPostgresql
Masters: [ pm01 ]
Slaves: [ pm02 ]
Clone Set: clnPingCheck
Started: [ pm01 pm02 ]
Node Attributes:
* Node pm01:
+ default_ping_set : 100
+ master-postgresql:0 : 1000
+ pgsql-data-status : LATEST
+ pgsql-master-baseline : 0000000059000078
+ pgsql-status : PRI
* Node pm02:
+ default_ping_set : 100
+ master-postgresql:1 : 100
+ pgsql-data-status : STREAMING|SYNC
+ pgsql-status : HS:sync
Migration summary:
* Node pm02:
* Node pm01:
- sync data and put pm01's postgresql.conf
(on pm01)
pm01# psql -h 192.168.3.2 -U postgres -c "SELECT pg_start_backup('label',true)"
pm01# rsync -avr --delete 192.168.3.2:/var/lib/pgsql/9.1/data/ /var/lib/pgsql/9.1/data/
pm01# psql -h 192.168.3.2 -U postgres -c "SELECT pg_stop_backup()"
pm01# su - postgres
pm01$ cp /path_to_backup/postgresql.conf /var/lib/pgsql/9.1/data/postgresql.conf
- delete PGSQL.lock if exist on pm01
pm01# rm /var/lib/pgsql/tmp/PGSQL.lock - cleanup failcount
pm01# crm resource cleanup msPostgresql cent01- wait to be Slave
- pgsql-status becomes "HS:sync" and pgsql-data-status becomes "STREAMING|SYNC"
(display example)
# crm_mon -Af
============
Last updated: Wed Feb 15 15:15:15 2012
Stack: Heartbeat
Current DC: pm02 (22222222-2222-2222-2222-222222222222) - partition with quorum
Version: 1.0.11-1554a83db0d3c3e546cfd3aaff6af1184f79ee87
2 Nodes configured, unknown expected votes
3 Resources configured.
============
Online: [ pm01 pm02 ]
Resource Group: master-group
vip-master (ocf::heartbeat:IPaddr2): Started pm02
vip-rep (ocf::heartbeat:IPaddr2): Started pm02
Master/Slave Set: msPostgresql
Masters: [ pm02 ]
Slaves: [ pm01 ]
Clone Set: clnPingCheck
Started: [ pm01 pm02 ]
Node Attributes:
* Node pm01:
+ default_ping_set : 100
+ master-postgresql:0 : 100
+ pgsql-data-status : STREAMING|SYNC
+ pgsql-status : HS:sync
* Node pm02:
+ default_ping_set : 100
+ master-postgresql:1 : 1000
+ pgsql-data-status : LATEST
+ pgsql-master-baseline : 0000000058000230
+ pgsql-status : PRI
Migration summary:
* Node pm02:
* Node pm01: