On Wed, Mar 2, 2011 at 6:22 AM, Simon Riggs <simon@2ndquadrant.com> wrote:
> The WALSender deliberately does *not* wake waiting users if the standby
> disconnects. Doing so would break the whole reason for having sync rep
> in the first place. What we do is allow a potential standby to takeover
> the role of sync standby, if one is available. Or the failing standby
> can reconnect and then release waiters.
If the transaction would have been allowed to commit without waiting
had the standby not been connected in the first place, then presumably
it should also be allowed to commit if the standby disconnects later,
too. Otherwise, it doesn't seem very consistent. A commit should
either wait for a disconnected standby to reconnect, or it should not
wait. It shouldn't wait in some situations but not others, I think.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company