I am not sure if I am hitting a bug… I am running 9.2.4
I created cascade standby as follows:
Master->cascading-standby->standby
I configured streaming replication. At first I thought everything worked. I created some data in master and went to cascading-standby and from there to standby. So, I thought it worked.
And then, I shutdown standby (the last one in the chain) and created data load in master for several hours. Everything got replicated to cascading-standby. However, when I restarted the standby (the one I shut down before), I got an error saying
“[2014-04-02 12:55:49 CDT][][]FATAL: could not receive data from WAL stream: FATAL: requested WAL segment 000000010000001C00000013 has already been removed”
So, everything worked in standby as long as the logs were available in pg_xlog area of cascading-standby. Once they are removed from pg_xlog, the streaming broke (this is expected).
Is there a way I can make cascading-standby send every log it gets from master down to standby server before it gets cleaned up by archive_cleanup_command? Or some other way??? (I don’t want master send logs to both slaves. That will be defeating the purpose of cascade set up.)
Thanks for your help.
Murthy