On 2/21/17 12:09 AM, Steven Chang wrote:
check this, his 2nd part introduces Stream Replication Implementation and tell you px_log retention and wal archive related parameters.
I'm interested in why I might observed a dramatic difference between pg_xlog files and the archive target files. Other dbs I see have them pretty much one-to-one, but not in this case here. What might cause the large variation? I got only a few pg_xlogs files each minute and 42 archive files/min.