Re: timeout on lock feature

Tom Lane wrote in message <4982(dot)987184866(at)sss(dot)pgh(dot)pa(dot)us>...
>Bruce Momjian <pgman(at)candle(dot)pha(dot)pa(dot)us> writes:
>> I was thinking SET because UPDATE does an auto-lock.
>
>Not to mention a ton of implicit locks acquired on various system tables
>during parsing/planning. You really want auto timeout on all of those?
>I sure don't.

*****************************************
YES, I DO! My transaction can't wait.
If parser on planner is blocked, then i want to abort my transaction.
*****************************************

>
>The appropriate way to do this given a LOCK TABLE option would be like
>
> BEGIN;
> LOCK TABLE foo IN ROW EXCLUSIVE MODE WITH TIMEOUT n;
> UPDATE foo SET ...;
> COMMIT;
>

*****************************************
With this solution, some server processes can block me!
*****************************************

>which restricts the scope of the timeout behavior to just the specific
>lock that the user is thinking of, and doesn't risk breaking fundamental
>system operations.
>
> regards, tom lane

*****************************************
This is real problem, but i think other postgres modules are ready for my
solution
(because it is extension to deadlock detection mechanism)
*****************************************

>
>---------------------------(end of broadcast)---------------------------
>TIP 6: Have you searched our list archives?
>
>http://www.postgresql.org/search.mpl