Alvaro Herrera <alvherre@2ndquadrant.com> writes:
> Tom Lane wrote:
>> After some discussion, the core committee has concluded that we should go
>> ahead with the already-wrapped releases. 9.2.6 and below are good anyway,
>> and despite this issue 9.3.2 is an improvement over 9.3.1. We'll plan to
>> do a 9.3.3 as soon as the multixact situation can be straightened out;
>> but let's learn from experience and not try to fix it in a panic.
> I would suggest we include this one fix in 9.3.2a. This bug is more
> serious than the others because it happens because of checking HTSU on a
> tuple containing running locker-only transactions and an aborted update.
The effect is just that the lockers could lose their locks early, right?
While that's annoying, it's not *directly* a data corruption problem.
And I've lost any enthusiasm I might've had for quick fixes in this area.
I think it'd be better to wait a few days, think this over, and get the
other problem fixed as well.
In any case, I think we're already past the point where we could re-wrap
9.3.2; the tarballs have been in the hands of packagers for > 24 hours.
We'd have to call it 9.3.3.
regards, tom lane
