| From: | Robert Haas <robertmhaas(at)gmail(dot)com> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: perform_spin_delay() vs wait events |
| Date: | 2022-11-21 15:35:16 |
| Message-ID: | CA+TgmoYNsNVcifkEEOGh4vyv0sm-6RBkGQ3Vn0ihpdqTNn2=mA@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Sun, Nov 20, 2022 at 6:10 PM Andres Freund <andres(at)anarazel(dot)de> wrote:
> I was wondering about that too - but decided against it because it would only
> show a single wait event. And wouldn't really describe spinlocks as a whole,
> just the "extreme" delays. If we wanted to report the spin waits more
> granular, we'd presumably have to fit the wait events into the lwlock, buffers
> and some new category where we name individual spinlocks.
>
> But I guess a single spinlock wait event type with an ExponentialBackoff wait
> event or such wouldn't be too bad.
Oh, hmm. I guess it is actually bracketing a timed wait, now that I
look closer at what you did. So perhaps your first idea was best after
all.
--
Robert Haas
EDB: http://www.enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Robert Haas | 2022-11-21 15:35:46 | Re: Damage control for planner's get_actual_variable_endpoint() runaway |
| Previous Message | Tom Lane | 2022-11-21 15:32:28 | Re: Damage control for planner's get_actual_variable_endpoint() runaway |