i have a pgpool2 3.1.3 installed with 2 postgresql 9.1.3 backend configured as master/slave with streaming replication.
If the master fails everything is ok, the slave takes over and becomes the new master.
The problem is that if i want to rejoin the old master in the cluster this is added as master also instead of slave.
I use pgpooladmin Version 3.1.1.
any idea ?
Asked
Active
Viewed 346 times
0
Videanu Adrian
- 972
- 3
- 16
- 41
-
Are you using pgpool for the replication or the streaming replication built in to PostgreSQL itself? (Or something else, like slony, etc.?) – kgrittn Apr 19 '12 at 12:25
-
i was using streaming replication build in postgres. – Videanu Adrian Apr 20 '12 at 14:38
2 Answers
0
i found the problem. When i first created the streaming replication i configured postgresql as in streaming replication tutorial. Here the slave and master had different settings activated in config file, and this was the problem.
After i read one more time the pgpool2 multi server tutorial i found that in the first stage both servers have the same setup, and the slave would be started through pgpool.
After i corrected that everything is ok.
Videanu Adrian
- 972
- 3
- 16
- 41
0
If you have successfully attached old master, then detaching new master wouldn't trigger old master get promoted again?
http://www.pgpool.net/docs/latest/pgpool-en.html#pcp_detach_node
Ochko
- 121
- 1
- 5