On Thu, Jun 29, 2017 at 4:27 PM, Thomas Munro
<thomas.munro@enterprisedb.com> wrote:
> /*
> * Acquiring the lock is not needed, the latch ensures proper
> * barriers. If it looks like we're done, we must really be done,
> * because once walsender changes the state to SYNC_REP_WAIT_COMPLETE,
> * it will never update it again, so we can't be seeing a stale value
> * in that case.
> */
Yeah, counting on the latch for free barriers doesn't work if you
happen to see SYNC_REP_WAIT_COMPLETE first time through the loop, or
if you see it after a spurious signal woke you and then it's
immediately set to SYNC_REP_WAIT_COMPLETE. In those cases, the
following Assert statement is making an assertion about cache
coherency that doesn't work even on a friendly TSO system.
Can you reproduce the problem with this experimental patch applied?
--
Thomas Munro
http://www.enterprisedb.com
--
Sent via pgsql-bugs mailing list (pgsql-bugs@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-bugs