Hi,
On 2022-02-02 13:59:56 +1300, Thomas Munro wrote:
> Seen again today on prairiedog. Erm, scratch that idea, HS feedback
> interferes with test results.
It'd not be sufficient anyway, I think. E.g. autovacuum truncating a table
would not be prevented by hs_f I think?
> I guess max_standby_streaming_delay
> should be increased to 'forever', like in the attached
Seems reasonable.
> , since pg_dump runs for a very long time on prairiedog:
> 2022-02-01 04:47:59.294 EST [3670:15] 027_stream_regress.pl LOG:
> statement: SET TRANSACTION ISOLATION LEVEL REPEATABLE READ, READ ONLY
> ...
> 2022-02-01 04:49:09.881 EST [3683:2585] 027_stream_regress.pl ERROR:
> canceling statement due to conflict with recovery
That, uh, seems slow. Is it perhaps waiting for a lock? Seems
Cluster.pm::init() should add at least log_lock_waits...
Greetings,
Andres Freund