On Mon, Feb 27, 2017 at 8:07 AM, Amit Kapila wrote:
>> My guess is that if we apply the fix I suggested above, it'll be good
>> enough. If that turns out not to be true, then I guess we'll have to
>> deal with that, but why not do the easy thing first?
>
> Okay, that is also a sensible approach.
On Sun, Feb 26, 2017 at 11:15 PM, Robert Haas wrote:
> On Mon, Feb 20, 2017 at 7:52 AM, Amit Kapila wrote:
>
>> The main point is if we
>> keep any loose end in this area, then there is a chance that the
>> regression test select_parallel can still fail, if not now, then in
>> future. Another wa
On Mon, Feb 20, 2017 at 7:52 AM, Amit Kapila wrote:
> On Sun, Feb 19, 2017 at 8:32 PM, Robert Haas wrote:
>> On Sun, Feb 19, 2017 at 6:50 PM, Amit Kapila wrote:
>>> To close the remaining gap, don't you think we can check slot->in_use
>>> flag when generation number for handle and slot are same.
On Sun, Feb 19, 2017 at 8:32 PM, Robert Haas wrote:
> On Sun, Feb 19, 2017 at 6:50 PM, Amit Kapila wrote:
>> To close the remaining gap, don't you think we can check slot->in_use
>> flag when generation number for handle and slot are same.
>
> That doesn't completely fix it either, because
> Forg
On Sun, Feb 19, 2017 at 6:50 PM, Amit Kapila wrote:
> To close the remaining gap, don't you think we can check slot->in_use
> flag when generation number for handle and slot are same.
That doesn't completely fix it either, because
ForgetBackgroundWorker() also does
BackgroundWorkerData->parallel_
On Sun, Feb 19, 2017 at 5:54 PM, Robert Haas wrote:
> On Sun, Feb 19, 2017 at 2:17 PM, Robert Haas wrote:
>> Such a change can be made, but as I pointed out in the part you didn't
>> quote, there are reasons to wonder whether that will be a constructive
>> change in real life even if it's better
On Sun, Feb 19, 2017 at 5:54 PM, Robert Haas wrote:
> However, it looks like there's a race condition here, because the slot
> doesn't get freed up at the same time that the PID gets set to 0.
> That actually happens later, when the postmaster calls
> maybe_start_bgworker() or DetermineSleepTime()
On Sun, Feb 19, 2017 at 2:17 PM, Robert Haas wrote:
> Such a change can be made, but as I pointed out in the part you didn't
> quote, there are reasons to wonder whether that will be a constructive
> change in real life even if it's better for the regression tests.
> Optimizing PostgreSQL for the
On Sat, Feb 18, 2017 at 11:53 PM, Tom Lane wrote:
>> It's a little surprising to me that the delay we're seeing here is
>> significant, because the death of a child should cause an immediate
>> SIGCHLD, resulting in a call to reaper(), resulting in a call to
>> waitpid(). Why's that not working?
Robert Haas writes:
> On Fri, Feb 17, 2017 at 9:57 PM, Amit Kapila wrote:
>>> That seems like a seriously broken design to me, first because it can make
>>> for a significant delay in the slots becoming available (which is what's
>>> evidently causing these regression failures), and second becaus
On Fri, Feb 17, 2017 at 9:57 PM, Amit Kapila wrote:
>> That seems like a seriously broken design to me, first because it can make
>> for a significant delay in the slots becoming available (which is what's
>> evidently causing these regression failures), and second because it's
>> simply bad desig
On Fri, Feb 17, 2017 at 9:15 PM, Tom Lane wrote:
> Amit Kapila writes:
>> On Fri, Feb 17, 2017 at 11:22 AM, Tom Lane wrote:
>>> In short, it looks to me like ExecShutdownGatherWorkers doesn't actually
>>> wait for parallel workers to finish (as its comment suggests is
>>> necessary), so that on
Amit Kapila writes:
> On Fri, Feb 17, 2017 at 11:22 AM, Tom Lane wrote:
>> In short, it looks to me like ExecShutdownGatherWorkers doesn't actually
>> wait for parallel workers to finish (as its comment suggests is
>> necessary), so that on not-too-speedy machines the worker slots may all
>> stil
On Fri, Feb 17, 2017 at 11:22 AM, Tom Lane wrote:
> Buildfarm members gaur, pademelon, and gharial have all recently shown
> failures like this:
>
> ***
> /home/bfarm/bf-data/HEAD/pgsql.build/src/test/regress/expected/select_parallel.out
> Thu Feb 16 20:35:14 2017
> ---
> /home/bfarm/bf-data/H
Buildfarm members gaur, pademelon, and gharial have all recently shown
failures like this:
***
/home/bfarm/bf-data/HEAD/pgsql.build/src/test/regress/expected/select_parallel.out
Thu Feb 16 20:35:14 2017
---
/home/bfarm/bf-data/HEAD/pgsql.build/src/test/regress/results/select_parallel.out
Th
15 matches
Mail list logo
