Search Postgresql Archives

Re: Strange issue with unique index

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



<rstander@xxxxxxxxx> writes:
> I've run into a strange issue with a unique index that I'm struggling to
> understand. I've extracted the basic info to reproduce this below.
> ...
> This will now block until session 2 is complete. I don't understand why this
> would block. I do know it's that unique index causing the issue, but I need
> the unique index in place.

No, it's not about the unique index.  It's about the foreign key
constraint --- if you remove that, there is no blockage.  The reason
why that's happening is that the insertions of dependent child rows
acquire row locks on the FK-referenced tuple, to prevent that row from
going away before the insertions commit.  So when you then decide to
UPDATE the referenced row, that blocks on the other session's row lock.

I had an idea that we were bright enough to not block if the UPDATE
doesn't change the column(s) involved in the FK, but either I'm wrong
or that's not working in this example for some reason.  It might be
that the fact that session 1 itself also holds such a row lock is
complicating matters.

You can make things a little better, at the cost of more overhead,
by declaring the FK as DEFERRABLE INITIALLY DEFERRED.

			regards, tom lane





[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Postgresql Jobs]     [Postgresql Admin]     [Postgresql Performance]     [Linux Clusters]     [PHP Home]     [PHP on Windows]     [Kernel Newbies]     [PHP Classes]     [PHP Databases]     [Postgresql & PHP]     [Yosemite]

  Powered by Linux