On 31 October 2014 19:36, Robert Haas <robertmhaas@gmail.com> wrote:
>> It's an obscure case and its not the only solution either.
>
> I don't think that's an obscure situation at all. Do you really think
> a patch that could cause an attempt to VACUUM FULL a system catalog to
> suffer an undetected deadlock meets this community's quality
> standards? Because that's what we're talking about.
Nobody has said that allowing undetected deadlocks is acceptable, so
your other comments are void.
I've suggested *stricter* locking, which would obviously allow
deadlock detection. You recognised that by claiming that the locking I
had proposed was actually too strict, which is where the above example
came from.
Yes, I have proposed stricter locking, but as explained, the only
things this would slow down are catalog VAC FULLs, which are already a
problem.
-- Simon Riggs http://www.2ndQuadrant.com/PostgreSQL Development, 24x7 Support, Training & Services