Horton

Re: Add an optional timeout clause to isolationtester step.

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Julien Rouhaud <rjuju123@gmail.com>
Cc: Alvaro Herrera <alvherre@2ndquadrant.com>, Michael Paquier <michael@paquier.xyz>, Andres Freund <andres@anarazel.de>, pgsql-hackers <pgsql-hackers@postgresql.org>
Date: 2020-03-13T16:58:25Z
Lists: pgsql-hackers 
Julien Rouhaud <rjuju123@gmail.com> writes:
> It seems that for all the possibly interesting cases, what we want to wait on
> is an heavyweight lock, which is already what isolationtester detects.  Maybe
> we could simply implement something like

> step "<name>" [ WAIT UNTIL BLOCKED ] { <SQL> }

> without any change to the blocking detection function?

Um, isn't that the existing built-in behavior?

I could actually imagine some uses for the reverse option, *don't* wait
for it to become blocked but just immediately continue with issuing
the next step.

			regards, tom lane

Commits

  1. Prevent reindex of invalid indexes on TOAST tables

  2. Fix more issues with dependency handling at swap phase of REINDEX CONCURRENTLY