On Fri, Dec 24, 2021 at 08:04:08PM +0530, Bharath Rupireddy wrote: > On Fri, Dec 24, 2021 at 7:19 PM Justin Pryzby <pry...@telsasoft.com> wrote: > > On Sun, Dec 12, 2021 at 06:08:05PM +0530, Bharath Rupireddy wrote: > > > Another idea could be to use the infrastructure laid out by the commit > > > 9ce346e [1]. With ereport_startup_progress, we can emit the LOGs(of > > > current recovering WAL file) for every log_startup_progress_interval > > > seconds/milliseconds. > > > > > > One problem is that ereport_startup_progress doesn't work on > > > StandbyMode, maybe we can remove this restriction unless we have a > > > major reason for not allowing it on the standby. > > > /* Prepare to report progress of the redo phase. */ > > > if (!StandbyMode) > > > begin_startup_progress_phase(); > > > > The relevant conversation starts here. > > https://www.postgresql.org/message-id/20210814214700.GO10479%40telsasoft.com > > > > There was a lot of confusion in that thread, though. > > > > The understanding was that it didn't make sense for that feature to > > continuously log messages on a standby (every 10sec by default). That seems > > like too much - the issue of a checkpointed logged every 5min was enough of > > a > > hurdle. > > > > If you're talking about a new feature that uses the infrastructre from 9ce3, > > but is controlled by a separate GUC like log_wal_traffic, that could be > > okay. > > Do you see any problems using the same GUC > log_startup_progress_interval and ereport_startup_progress API > introduced by 9ce346e? I prefer this instead of a new GUC. Thoughts?
I referenced the thread where we avoided writing progress logs during normal operation of a standby, since the logs would be both useless and excessive. If your patch were to use the same infrastructure, you'd still want to avoid disturbing that behavior and writing useless logs. -- Justin