On 17/09/2020 13:31, Thomas Munro wrote:
> On Thu, Sep 17, 2020 at 10:19 PM Heikki Linnakangas <hlinnaka@iki.fi> wrote:
>> If you put the counter in HandleStartupProcInterrupts(), it could be a
>> long wait if the startup process is e.g. waiting for WAL to arrive in
>> the loop in WaitForWALToBecomeAvailable(), or in recoveryPausesHere().
>> My original patch only reduced the frequency in the WAL redo loop, when
>> you're actively replaying records.
>
> Oh, I checked that WaitForWALToBecomeAvailable() already handled
> postmaster death via events rather than polling, with
> WL_EXIT_ON_PM_DEATH, but I hadn't clocked that recoveryPausesHere()
> uses pg_usleep() and polling. Hmm. Perhaps we should change that
> instead? The reason I did it that way is that I didn't want to make
> the new ProcSignalBarrierPending handler less reactive.
Hmm, so for speedy response to postmaster death, you're relying on the
loops to have other postmaster-death checks besides
HandleStartupProcInterrupts(), in the form of WL_EXIT_ON_PM_DEATH. That
seems a bit fragile, at the very least it needs a comment in
HandleStartupProcInterrupts() to call it out.
Note that there's one more loop in ShutdownWalRcv() that uses pg_usleep().
- Heikki
