We had two servers set up with streaming replication. I believe when we did a full vacuum analyze to the database on the master, it caused the slave to not be able to catch up (wal_keep_segments = 32, but still not enough and had archive off). I am trying to re-set up the slave to streaming replication. Below are the steps I took (attempted a couple times now). Does anyone have any idea what I am doing wrong? The service is unable to start after I copy and paste the files over.
- Uninstalled and re-installed postgres on slave and master (v9.1.10)
- Stopped slave postgres service and added files and settings for streaming replication to the slave (postgresql.cof, recovery.conf - which all worked in the past set up)
- Loaded the master with a pg_dumpall backup file.
- Stopped the master postgres service.
- Copied the data directory from master to slave minus pg_xlog folder, postgresql.conf and postgresql.pid files using rsync
- Attempted to start postgres on the slave and it is unable to start
- Postgres log:
- CDT LOG: entering standby mode
- CDT FATAL: the database is starting up (multiple times)
- CDT FATAL: could not connect to the primary server... (assume this is ok because master has not been started yet)
- Event Viewer, Administrative Events
- Timed out waiting for server start up
Any help, thoughts, comments, tips, etc would be greatly appreciated.
Thanks!
Dara