Re: Excessive PostmasterIsAlive calls slow down WAL redo

On April 9, 2018 6:57:23 PM PDT, Alvaro Herrera <alvherre@alvh.no-ip.org> wrote:
>Andres Freund wrote:
>> 
>> On April 9, 2018 6:31:07 PM PDT, Alvaro Herrera
><alvherre@alvh.no-ip.org> wrote:
>
>> >Would it work to use this second pipe, to which each child writes a
>> >byte that postmaster never reads, and then rely on SIGPIPE when
>> >postmaster dies?  Then we never need to do a syscall.
>> 
>> I'm not following, could you expand on what you're suggesting?  Note
>> that you do not get SIGPIPE for already buffered writes.  Which
>> syscall can we avoid?
>
>Ah.  I was thinking we'd get SIGPIPE from the byte sent at the start,
>as
>soon as the kernel saw that postmaster abandoned the fd by dying.
>Scratch that then.

Had the same idea, but unfortunately reality, in the form of a test program, cured me of my hope ;)
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.