Re: Allow interrupts on waiting standby

On Fri, Jan 27, 2017 at 9:23 PM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> On 27 January 2017 at 01:35, Michael Paquier <michael(dot)paquier(at)gmail(dot)com> wrote:
>> On Fri, Jan 27, 2017 at 4:36 AM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
>>> On 26 January 2017 at 19:20, Andres Freund <andres(at)anarazel(dot)de> wrote:
>> A deeper fix for HEAD proves to not be that complicated. Please see
>> the attached. The other two calls of pg_usleep() in standby.c are
>> waiting with 5ms and 10ms, it is not worth switching them to a latch.
>
> So you think 2 calls to pg_usleep() can stay; my opinion is 3 can stay.

This patch replaces one call of pg_usleep() where the wait can go up
to 1s, this is largely noticeable by the user. The two others sleep
for a maximum of 10ms. Even if a latch is used the code path is going
to exit immediately anyway. That's why you added a call to
CHECK_FOR_INTERRUPTS only here, no?

> I'm not clear why this particular call is worthy, while dozens of
> calls in other modules remain unchanged. This seems like a code issue
> rather than anything to do with Hot Standby in particular, so it
> should be another thread.

Some should switch to Latch.

> Doesn't seem important compared to other
> things for this release I should work on.

That's your call.

> Please add to the next CF so it gets proper review.

Sure.
--
Michael