On 2018-Oct-14, Andres Freund wrote:
> On 2018-10-14 13:26:24 +0000, Michael Paquier wrote:
> > Avoid duplicate XIDs at recovery when building initial snapshot
> I'm unhappy this approach was taken over objections. Without a real
> warning. Even leaving the crummyness aside, did you check other users
> of XLOG_RUNNING_XACTS, e.g. logical decoding?
Mumble. Is there a real problem here -- I mean, did this break logical
decoding? Maybe we need some more tests to ensure this stuff works
sanely for logical decoding.
FWIW and not directly related, I recently became aware (because of a
customer question) that txid_current_snapshot() is oblivious of
XLOG_RUNNING_XACTS in a standby. AFAICS it's not a really serious
concern, but it was surprising.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services