On Fri, Jul 4, 2014 at 7:29 PM, MauMau <maumau...@gmail.com> wrote: > > Hello, > > "When requesting synchronous replication, each commit of a write transaction will wait until confirmation is received that the commit has been written to the transaction log on disk of both the primary and standby server. > ... > Read only transactions and transaction rollbacks need not wait for replies from standby servers. Subtransaction commits do not wait for responses from standby servers, only top-level commits." > > > [Hypothesis] > Why does the connection processing emit WAL? > > Probably, it did page-at-a-time vacuum during access to pg_database and pg_authid for client authentication. src/backend/access/heap/README.HOT describes:
I agree with your analysis that it can happen during connection attempt. > But the customer could not reproduce the problem when he performed the same archive recovery from the same base backup again. Why? I guess the autovacuum daemon vacuumed the system catalogs before he attempted to connect to the database. > > Is this correct? One way to confirm could be to perform the archive recovery by disabling autovacuum. > > [How to fix] > Of course, adding "-o '-c synchronous_commit=local'" or "-o '-c synchronous_standby_names='" to pg_ctl start in the recovery script would prevent the problem. > > But isn't there anything to fix in PostgreSQL? I think the doc needs improvement so that users won't misunderstand that only write transactions would block at commit. I also think at the very least we should update docs even if we don't have any solution for this case. Another thing which I am wondering about is can't the same happen even for Read Only transaction (incase someone does Select which prunes the page). With Regards, Amit Kapila. EnterpriseDB: http://www.enterprisedb.com