PostgreSQL 一主两备节点（两备节点为同步节点）故障恢复
2016-03-10 19:53 DataBases 阅读(3256) 评论(0) 编辑 收藏 举报PostgreSQL 同步复制及故障恢复
10.2.208.10:node1:master
10.2.208.11:node2:standby1 同步
10.2.208.12:node3:standby2 同步
10.2.208.13:node4:备用机
node1，node2,node3,node4均安装PostgreSQL数据库，master初始化数据库，standby1与standby2从master pg_basebackup。
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica -d postgres
psql (9.5.1)
Type "help" for help.
postgres=# LOG: standby "node2" is now the synchronous standby with priority 1
postgres=# select version();
version
---------------------------------------------------------------------------------------------------------
PostgreSQL 9.5.1 on x86_64-pc-linux-gnu, compiled by gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-9), 64-bit
(1 row)
postgres=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
f
(1 row)
postgres=# \x
Expanded display is on.
postgres=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 14818
usesysid | 10
usename | replica
application_name | node2
client_addr | 10.2.208.11
client_hostname |
client_port | 44530
backend_start | 2016-03-10 11:16:10.850257+08
backend_xmin |
state | streaming
sent_location | 0/4000220
write_location | 0/4000220
flush_location | 0/4000220
replay_location | 0/4000220
sync_priority | 1
sync_state | sync
-[ RECORD 2 ]----+------------------------------
pid | 14820
usesysid | 10
usename | replica
application_name | node3
client_addr | 10.2.208.12
client_hostname |
client_port | 32839
backend_start | 2016-03-10 11:16:28.010007+08
backend_xmin |
state | streaming
sent_location | 0/4000220
write_location | 0/4000220
flush_location | 0/4000220
replay_location | 0/4000220
sync_priority | 2
sync_state | potential
[replica@node2 bin]$ ./psql -h localhost -p 6432 -U replica -d postgres
psql (9.5.1)
Type "help" for help.
postgres=# select version();
version
---------------------------------------------------------------------------------------------------------
PostgreSQL 9.5.1 on x86_64-pc-linux-gnu, compiled by gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-9), 64-bit
(1 row)
postgres=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
postgres=# \x
Expanded display is on.
postgres=# select * from pg_stat_replication ;
(0 rows)
[replica@node3 bin]$ ./psql -h localhost -p 6432 -U replica -d postgres
psql (9.5.1)
Type "help" for help.
postgres=# select version();
version
---------------------------------------------------------------------------------------------------------
PostgreSQL 9.5.1 on x86_64-pc-linux-gnu, compiled by gcc (GCC) 4.8.3 20140911 (Red Hat 4.8.3-9), 64-bit
(1 row)
postgres=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
postgres=# \x
Expanded display is on.
postgres=# select * from pg_stat_replication ;
(0 rows)
[replica@node1 ~]$ ps -ef|grep postgres
replica 14731 1 0 11:10 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 14733 14731 0 11:10 ? 00:00:00 postgres: checkpointer process
replica 14734 14731 0 11:10 ? 00:00:00 postgres: writer process
replica 14735 14731 0 11:10 ? 00:00:00 postgres: wal writer process
replica 14736 14731 0 11:10 ? 00:00:00 postgres: autovacuum launcher process
replica 14737 14731 0 11:10 ? 00:00:00 postgres: archiver process last was 000000010000000000000003.00000028.backup
replica 14738 14731 0 11:10 ? 00:00:00 postgres: stats collector process
replica 14749 13945 0 11:10 pts/2 00:00:00 ./psql -h localhost -p 6432 -U replica -d postgres
replica 14750 14731 0 11:10 ? 00:00:00 postgres: replica postgres ::1(43486) idle
replica 14818 14731 0 11:16 ? 00:00:00 postgres: wal sender process replica 10.2.208.11(44530) streaming 0/4000300
replica 14820 14731 0 11:16 ? 00:00:00 postgres: wal sender process replica 10.2.208.12(32839) streaming 0/4000300
replica 15020 14302 0 11:29 pts/0 00:00:00 grep --color=auto postgres
[replica@node2 ~]$ ps -ef|grep postgres
replica 27194 1 0 11:16 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 27195 27194 0 11:16 ? 00:00:00 postgres: startup process recovering 000000010000000000000004
replica 27196 27194 0 11:16 ? 00:00:00 postgres: checkpointer process
replica 27197 27194 0 11:16 ? 00:00:00 postgres: writer process
replica 27198 27194 0 11:16 ? 00:00:00 postgres: stats collector process
replica 27199 27194 0 11:16 ? 00:00:00 postgres: wal receiver process streaming 0/4000300
replica 27216 26525 0 11:18 pts/2 00:00:00 ./psql -h localhost -p 6432 -U replica -d postgres
replica 27217 27194 0 11:18 ? 00:00:00 postgres: replica postgres ::1(43081) idle
replica 27417 27091 0 11:30 pts/0 00:00:00 grep --color=auto postgres
[replica@node3 ~]$ ps -ef|grep postgres
replica 20747 1 0 11:16 pts/1 00:00:00 /home/replica/bin/postgres -D ../data
replica 20748 20747 0 11:16 ? 00:00:00 postgres: startup process recovering 000000010000000000000004
replica 20749 20747 0 11:16 ? 00:00:00 postgres: checkpointer process
replica 20750 20747 0 11:16 ? 00:00:00 postgres: writer process
replica 20751 20747 0 11:16 ? 00:00:00 postgres: stats collector process
replica 20752 20747 0 11:16 ? 00:00:00 postgres: wal receiver process streaming 0/4000300
replica 20769 20161 0 11:18 pts/1 00:00:00 ./psql -h localhost -p 6432 -U replica -d postgres
replica 20770 20747 0 11:18 ? 00:00:00 postgres: replica postgres ::1(58250) idle
replica 20936 20653 0 11:31 pts/2 00:00:00 grep --color=auto postgres
模拟同步节点Standby1 (node2)故障
1 停止Standby1的数据库服务
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica -d replica
psql (9.5.1)
Type "help" for help.
replica=# \d+
List of relations
Schema | Name | Type | Owner | Size | Description
--------+------------------+-------+---------+---------+-------------
public | pgbench_accounts | table | replica | 3651 MB |
public | pgbench_branches | table | replica | 48 kB |
public | pgbench_history | table | replica | 0 bytes |
public | pgbench_tellers | table | replica | 160 kB |
(4 rows)
replica=# LOG: standby "node3" is now the synchronous standby with priority 2
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
f
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 14820
usesysid | 10
usename | replica
application_name | node3
client_addr | 10.2.208.12
client_hostname |
client_port | 32839
backend_start | 2016-03-10 11:16:28.010007+08
backend_xmin |
state | streaming
sent_location | 0/DEF730B8
write_location | 0/DEF730B8
flush_location | 0/DEF730B8
replay_location | 0/DEF730B8
sync_priority | 2
sync_state | sync
[replica@node1 ~]$ ps -ef|grep postgres
replica 14731 1 0 11:10 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 14733 14731 0 11:10 ? 00:00:00 postgres: checkpointer process
replica 14734 14731 0 11:10 ? 00:00:00 postgres: writer process
replica 14735 14731 0 11:10 ? 00:00:01 postgres: wal writer process
replica 14736 14731 0 11:10 ? 00:00:00 postgres: autovacuum launcher process
replica 14737 14731 0 11:10 ? 00:00:00 postgres: archiver process last was 0000000100000000000000DD
replica 14738 14731 0 11:10 ? 00:00:00 postgres: stats collector process
replica 14820 14731 0 11:16 ? 00:00:02 postgres: wal sender process replica 10.2.208.12(32839) streaming 0/DEF73198
replica 15378 14731 0 11:36 ? 00:00:00 postgres: replica replica ::1(43489) idle
replica 15541 14302 0 11:47 pts/0 00:00:00 grep --color=auto postgres
[replica@node3 bin]$ ./psql -h localhost -p 6432 -U replica -d postgres
psql (9.5.1)
Type "help" for help.
postgres=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
postgres=# \x
Expanded display is on.
postgres=# select * from pg_stat_replication ;
(0 rows)
[replica@node3 ~]$ ps -ef|grep postgres
replica 20747 1 0 11:16 pts/1 00:00:00 /home/replica/bin/postgres -D ../data
replica 20748 20747 0 11:16 ? 00:00:16 postgres: startup process recovering 0000000100000000000000DE
replica 20749 20747 0 11:16 ? 00:00:03 postgres: checkpointer process
replica 20750 20747 0 11:16 ? 00:00:00 postgres: writer process
replica 20751 20747 0 11:16 ? 00:00:00 postgres: stats collector process
replica 20752 20747 0 11:16 ? 00:00:10 postgres: wal receiver process streaming 0/DEF73198
replica 21127 20161 0 11:48 pts/1 00:00:00 ./psql -h localhost -p 6432 -U replica -d postgres
replica 21128 20747 0 11:48 ? 00:00:00 postgres: replica postgres ::1(58252) idle
replica 21196 20653 0 11:50 pts/2 00:00:00 grep --color=auto postgres
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica -d replica
psql (9.5.1)
Type "help" for help.
replica=# create table test_tb(id int);
CREATE TABLE
replica=# insert into test_tb values (1),(2),(3);
INSERT 0 3
replica=# select * from test_tb ;
id
----
1
2
3
(3 rows)
[replica@node3 bin]$ ./psql -h localhost -p 6432 -U replica -d replica
psql (9.5.1)
Type "help" for help.
replica=# select * from test_tb ;
id
----
1
2
3
(3 rows)
重新启动Standby1
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica -d replica
psql (9.5.1)
Type "help" for help.
replica=# select * from test_tb ;
id
----
1
2
3
(3 rows)
replica=# LOG: standby "node2" is now the synchronous standby with priority 1
replica=#
查看复制状态
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica -d replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
f
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 15705
usesysid | 10
usename | replica
application_name | node2
client_addr | 10.2.208.11
client_hostname |
client_port | 44533
backend_start | 2016-03-10 11:56:10.103849+08
backend_xmin |
state | streaming
sent_location | 0/DEF89618
write_location | 0/DEF89618
flush_location | 0/DEF89618
replay_location | 0/DEF89618
sync_priority | 1
sync_state | sync
-[ RECORD 2 ]----+------------------------------
pid | 14820
usesysid | 10
usename | replica
application_name | node3
client_addr | 10.2.208.12
client_hostname |
client_port | 32839
backend_start | 2016-03-10 11:16:28.010007+08
backend_xmin |
state | streaming
sent_location | 0/DEF89618
write_location | 0/DEF89618
flush_location | 0/DEF89618
replay_location | 0/DEF89618
sync_priority | 2
sync_state | potential
[replica@node1 ~]$ ps -ef|grep postgres
replica 14731 1 0 11:10 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 14733 14731 0 11:10 ? 00:00:00 postgres: checkpointer process
replica 14734 14731 0 11:10 ? 00:00:00 postgres: writer process
replica 14735 14731 0 11:10 ? 00:00:01 postgres: wal writer process
replica 14736 14731 0 11:10 ? 00:00:00 postgres: autovacuum launcher process
replica 14737 14731 0 11:10 ? 00:00:00 postgres: archiver process last was 0000000100000000000000DD
replica 14738 14731 0 11:10 ? 00:00:00 postgres: stats collector process
replica 14820 14731 0 11:16 ? 00:00:02 postgres: wal sender process replica 10.2.208.12(32839) streaming 0/DEF896F8
replica 15705 14731 0 11:56 ? 00:00:00 postgres: wal sender process replica 10.2.208.11(44533) streaming 0/DEF896F8
replica 15861 14731 0 12:01 ? 00:00:00 postgres: replica replica ::1(43491) idle
replica 15946 14302 0 12:07 pts/0 00:00:00 grep --color=auto postgres
[replica@node2 bin]$ ./psql -h localhost -p 6432 -U replica -d replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
(0 rows)
[replica@node2 ~]$ ps -ef|grep postgres
replica 27714 1 0 11:56 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 27715 27714 0 11:56 ? 00:00:00 postgres: startup process recovering 0000000100000000000000DE
replica 27716 27714 0 11:56 ? 00:00:00 postgres: checkpointer process
replica 27717 27714 0 11:56 ? 00:00:00 postgres: writer process
replica 27718 27714 0 11:56 ? 00:00:00 postgres: stats collector process
replica 27719 27714 0 11:56 ? 00:00:00 postgres: wal receiver process streaming 0/DEF896F8
replica 27908 27714 0 12:05 ? 00:00:00 postgres: replica replica ::1(43084) idle
replica 27942 27091 0 12:08 pts/0 00:00:00 grep --color=auto postgres
[replica@node3 bin]$ ./psql -h localhost -p 6432 -U replica -d replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
(0 rows)
[replica@node3 ~]$ ps -ef|grep postgres
replica 20747 1 0 11:16 pts/1 00:00:00 /home/replica/bin/postgres -D ../data
replica 20748 20747 0 11:16 ? 00:00:16 postgres: startup process recovering 0000000100000000000000DE
replica 20749 20747 0 11:16 ? 00:00:03 postgres: checkpointer process
replica 20750 20747 0 11:16 ? 00:00:00 postgres: writer process
replica 20751 20747 0 11:16 ? 00:00:00 postgres: stats collector process
replica 20752 20747 0 11:16 ? 00:00:11 postgres: wal receiver process streaming 0/DEF896F8
replica 21444 20747 0 12:06 ? 00:00:00 postgres: replica replica ::1(58254) idle
replica 21507 20653 0 12:10 pts/2 00:00:00 grep --color=auto postgres
模拟Standby2(node3)故障
[replica@node3 bin]$ ./pg_ctl -D ../data/ stop
查看状态
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
f
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 15705
usesysid | 10
usename | replica
application_name | node2
client_addr | 10.2.208.11
client_hostname |
client_port | 44533
backend_start | 2016-03-10 11:56:10.103849+08
backend_xmin |
state | streaming
sent_location | 0/DEF89A78
write_location | 0/DEF89A78
flush_location | 0/DEF89A78
replay_location | 0/DEF89A78
sync_priority | 1
sync_state | sync
[replica@node1 ~]$ ps -ef|grep postgres
replica 14731 1 0 11:10 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 14733 14731 0 11:10 ? 00:00:00 postgres: checkpointer process
replica 14734 14731 0 11:10 ? 00:00:00 postgres: writer process
replica 14735 14731 0 11:10 ? 00:00:01 postgres: wal writer process
replica 14736 14731 0 11:10 ? 00:00:00 postgres: autovacuum launcher process
replica 14737 14731 0 11:10 ? 00:00:00 postgres: archiver process last was 0000000100000000000000DD
replica 14738 14731 0 11:10 ? 00:00:00 postgres: stats collector process
replica 15705 14731 0 11:56 ? 00:00:00 postgres: wal sender process replica 10.2.208.11(44533) streaming 0/DEF8A5F0
replica 16112 14731 0 12:17 ? 00:00:00 postgres: replica replica ::1(43492) idle
replica 16329 14731 0 12:22 ? 00:00:00 postgres: replica replica ::1(43493) idle
replica 16493 14731 0 12:27 ? 00:00:00 postgres: replica replica ::1(43494) idle
replica 17772 17733 0 13:34 pts/5 00:00:00 grep --color=auto postgres
[replica@node2 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
(0 rows)
[replica@node2 ~]$ ps -ef|grep postgres
replica 27714 1 0 11:56 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 27715 27714 0 11:56 ? 00:00:00 postgres: startup process recovering 0000000100000000000000DE
replica 27716 27714 0 11:56 ? 00:00:00 postgres: checkpointer process
replica 27717 27714 0 11:56 ? 00:00:00 postgres: writer process
replica 27718 27714 0 11:56 ? 00:00:00 postgres: stats collector process
replica 27719 27714 0 11:56 ? 00:00:01 postgres: wal receiver process streaming 0/DEF8A5F0
replica 27908 27714 0 12:05 ? 00:00:00 postgres: replica replica ::1(43084) idle
replica 28311 27714 0 12:28 ? 00:00:00 postgres: replica replica ::1(43085) idle
replica 29330 29291 0 13:36 pts/3 00:00:00 grep --color=auto postgres
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# create table test_tb2(id int);
CREATE TABLE
replica=# insert into test_tb2 values (1),(2),(3);
INSERT 0 3
[replica@node2 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# select * from test_tb2;
id
----
1
2
3
(3 rows)
重新启动Standby2(node3)
[replica@node3 bin]$ ./pg_ctl -D ../data/ start
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
f
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 15705
usesysid | 10
usename | replica
application_name | node2
client_addr | 10.2.208.11
client_hostname |
client_port | 44533
backend_start | 2016-03-10 11:56:10.103849+08
backend_xmin |
state | streaming
sent_location | 0/DEFA14B0
write_location | 0/DEFA14B0
flush_location | 0/DEFA14B0
replay_location | 0/DEFA14B0
sync_priority | 1
sync_state | sync
-[ RECORD 2 ]----+------------------------------
pid | 17908
usesysid | 10
usename | replica
application_name | node3
client_addr | 10.2.208.12
client_hostname |
client_port | 32845
backend_start | 2016-03-10 13:40:58.96505+08
backend_xmin |
state | streaming
sent_location | 0/DEFA14B0
write_location | 0/DEFA14B0
flush_location | 0/DEFA14B0
replay_location | 0/DEFA14B0
sync_priority | 2
sync_state | potential
[replica@node1 ~]$ ps -ef|grep postgres
replica 14731 1 0 11:10 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 14733 14731 0 11:10 ? 00:00:00 postgres: checkpointer process
replica 14734 14731 0 11:10 ? 00:00:00 postgres: writer process
replica 14735 14731 0 11:10 ? 00:00:01 postgres: wal writer process
replica 14736 14731 0 11:10 ? 00:00:00 postgres: autovacuum launcher process
replica 14737 14731 0 11:10 ? 00:00:00 postgres: archiver process last was 0000000100000000000000DD
replica 14738 14731 0 11:10 ? 00:00:00 postgres: stats collector process
replica 15705 14731 0 11:56 ? 00:00:00 postgres: wal sender process replica 10.2.208.11(44533) streaming 0/DEFA14B0
replica 16112 14731 0 12:17 ? 00:00:00 postgres: replica replica ::1(43492) idle
replica 16329 14731 0 12:22 ? 00:00:00 postgres: replica replica ::1(43493) idle
replica 17908 14731 0 13:40 ? 00:00:00 postgres: wal sender process replica 10.2.208.12(32845) streaming 0/DEFA14B0
replica 17913 14731 0 13:41 ? 00:00:00 postgres: replica replica ::1(43496) idle
replica 17967 17733 0 13:45 pts/5 00:00:00 grep --color=auto postgres
[replica@node2 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
(0 rows)
[replica@node2 ~]$ ps -ef|grep postgres
replica 27714 1 0 11:56 pts/2 00:00:00 /home/replica/bin/postgres -D ../data
replica 27715 27714 0 11:56 ? 00:00:00 postgres: startup process recovering 0000000100000000000000DE
replica 27716 27714 0 11:56 ? 00:00:00 postgres: checkpointer process
replica 27717 27714 0 11:56 ? 00:00:00 postgres: writer process
replica 27718 27714 0 11:56 ? 00:00:00 postgres: stats collector process
replica 27719 27714 0 11:56 ? 00:00:01 postgres: wal receiver process streaming 0/DEFA1590
replica 27908 27714 0 12:05 ? 00:00:00 postgres: replica replica ::1(43084) idle
replica 29467 27714 0 13:46 ? 00:00:00 postgres: replica replica ::1(43087) idle
replica 29478 29291 0 13:47 pts/3 00:00:00 grep --color=auto postgres
[replica@node3 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
t
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
(0 rows)
[replica@node3 ~]$ ps -ef|grep postgres
replica 22845 1 0 13:40 pts/1 00:00:00 /home/replica/bin/postgres -D ../data
replica 22846 22845 0 13:40 ? 00:00:00 postgres: startup process recovering 0000000100000000000000DE
replica 22847 22845 0 13:40 ? 00:00:00 postgres: checkpointer process
replica 22848 22845 0 13:40 ? 00:00:00 postgres: writer process
replica 22849 22845 0 13:40 ? 00:00:00 postgres: stats collector process
replica 22850 22845 0 13:40 ? 00:00:00 postgres: wal receiver process streaming 0/DEFA1590
replica 22919 22845 0 13:48 ? 00:00:00 postgres: replica replica ::1(58256) idle
replica 23029 22982 0 13:49 pts/2 00:00:00 grep --color=auto postgres
故障恢复处理
模拟Standby2故障
1 停止node3[10.2.208.12]数据库的服务
pg_ctl –D ../data stop –m fast
2 更改master[node1][10.2.208.10]postgresql.conf文件
synchronous_standby_names = ' node2,node4'
pg_ctl –D ../data reload
3 node4[10.2.208.13]从master pg_basebackup
mv recovery.done recovery.conf
primary_conninfo = 'host=10.2.208.10 port=6432 user=replica password=replica application_name=node4'
4 启动node4
pg_ctl –D ../data start
[replica@node1 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# select pg_is_in_recovery();
pg_is_in_recovery
-------------------
f
(1 row)
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 15705
usesysid | 10
usename | replica
application_name | node2
client_addr | 10.2.208.11
client_hostname |
client_port | 44533
backend_start | 2016-03-10 11:56:10.103849+08
backend_xmin |
state | streaming
sent_location | 0/E0000060
write_location | 0/E0000060
flush_location | 0/E0000060
replay_location | 0/E0000060
sync_priority | 1
sync_state | sync
-[ RECORD 2 ]----+------------------------------
pid | 19907
usesysid | 10
usename | replica
application_name | node4
client_addr | 10.2.208.13
client_hostname |
client_port | 34767
backend_start | 2016-03-10 15:08:29.425353+08
backend_xmin |
state | streaming
sent_location | 0/E0000060
write_location | 0/E0000060
flush_location | 0/E0000060
replay_location | 0/E0000060
sync_priority | 2
sync_state | potential
模拟Standby1故障
1 停止Standby1 的数据库服务
pg_ctl –D ../data stop –m fast
2更改master[10.2.208.10]的postgresql.conf
synchronous_standby_names = ' node3,node4'
pg_ctl –D ../data reload
3在node4[10.2.208.13] 从master pg_basebackup
mv recovery.done recovery.conf
primary_conninfo = 'host=10.2.208.10 port=6432 user=replica password=replica application_name=node4'
4 启动node4
pg_ctl –D ../data start
master
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 20668
usesysid | 10
usename | replica
application_name | node4
client_addr | 10.2.208.13
client_hostname |
client_port | 34770
backend_start | 2016-03-10 15:45:49.027116+08
backend_xmin |
state | streaming
sent_location | 0/E4000060
write_location | 0/E4000060
flush_location | 0/E4000060
replay_location | 0/E4000000
sync_priority | 2
sync_state | potential
-[ RECORD 2 ]----+------------------------------
pid | 20210
usesysid | 10
usename | replica
application_name | node3
client_addr | 10.2.208.12
client_hostname |
client_port | 32856
backend_start | 2016-03-10 15:18:01.507366+08
backend_xmin |
state | streaming
sent_location | 0/E4000060
write_location | 0/E4000060
flush_location | 0/E4000060
replay_location | 0/E4000060
sync_priority | 1
sync_state | sync
模拟主库故障
1 停止master[node1]的数据库服务
pg_ctl –D ../data stop –m fast
2更改Standby1[node2]的postgresql.conf文件
synchronous_standby_names = 'node3,node4'
pg_ctl –D ../data reload
pg_ctl –D ../data promote
3更改Standby2[node3]的recovery.conf文件
primary_conninfo = 'host=10.2.208.11 port=6432 user=replica password=replica application_name=node3'
pg_ctl –D ../data restart
在node2上查看复制状态
select * from pg_stat_replication;
[replica@node2 bin]$ ./psql -h localhost -p 6432 -U replica
psql (9.5.1)
Type "help" for help.
replica=# LOG: standby "node3" is now the synchronous standby with priority 1
replica=# \x
Expanded display is on.
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+-----------------------------
pid | 31996
usesysid | 10
usename | replica
application_name | node3
client_addr | 10.2.208.12
client_hostname |
client_port | 38750
backend_start | 2016-03-10 16:21:42.23142+08
backend_xmin |
state | streaming
sent_location | 0/E7000060
write_location | 0/E7000060
flush_location | 0/E7000060
replay_location | 0/E7000060
sync_priority | 1
sync_state | sync
4在node4[10.2.208.13] 从node2[10.2.208.11] pg_basebackup
mv recovery.done recovery.conf
primary_conninfo = 'host=10.2.208.11 port=6432 user=replica password=replica application_name=node4'
4 启动node4
pg_ctl –D ../data start
replica=# select * from pg_stat_replication ;
-[ RECORD 1 ]----+------------------------------
pid | 31996
usesysid | 10
usename | replica
application_name | node3
client_addr | 10.2.208.12
client_hostname |
client_port | 38750
backend_start | 2016-03-10 16:21:42.23142+08
backend_xmin |
state | streaming
sent_location | 0/E7000060
write_location | 0/E7000060
flush_location | 0/E7000060
replay_location | 0/E7000060
sync_priority | 1
sync_state | sync
-[ RECORD 2 ]----+------------------------------
pid | 32031
usesysid | 10
usename | replica
application_name | node4
client_addr | 10.2.208.13
client_hostname |
client_port | 53133
backend_start | 2016-03-10 16:25:27.541883+08
backend_xmin |
state | streaming
sent_location | 0/E7000060
write_location | 0/E7000060
flush_location | 0/E7000060
replay_location | 0/E7000000
sync_priority | 2
sync_state | potential
1主节点+1同步节点+1异步节点：缺点，同步节点故障，主节点将不提供写操作(insert,update)
三节点级联复制：1主节点+1上游节点(同步)+1下游节点：上游节点故障，主节点将不提供写操作(insert ,update)
同步复制中的recovery.conf 中的application_name与postgresql.conf中的synchronous_standby_names可以使用IP地址代替hostname
