I actually i am in a position to do rigorous stress testing because i have the machines at my disposal. I am sorry that i was not in touch with the lists for a prolong duration and could not followup much on the discussions while HS & SR were being developed.hence i am mostly relying on the published documents only. The wiki that i used does not say much about how to manage the wal files that are archived in primary and standby(possibly).
I am currently aiming to setup only SR between 2 servers only.
in primary server i have --------------------------------------------------------------------- wal_level = hot_standby max_wal_senders = 5 wal_keep_segments = 32 archive_mode = on archive_command = 'cp %p /mnt/disk1/pgdatadir/wal_archives/%f' -----------------------------------------------------------------------------
in slave sever i have additionally hot_standby = on # in postgresql.conf
and recovery.con in slave is: -------------- standby_mode = 'on' primary_conninfo = 'host=172.16.10.139 port=5432 user=postgres' trigger_file = '/mnt/disk1/pgdatadir/trigger' restore_command = 'cp /mnt/disk1/pgdatadir/wal_archives/%f "%p"' -----------------------
My question is in SR setup do i need to keep transferring the archived WAL files also ? when the slave requests WAL records from the master does the master consult the archived wal files also for sending the records ?
