On 01/03/2015 10:42 PM, Peter Geoghegan wrote:
On Sat, Jan 3, 2015 at 2:41 AM, Heikki Linnakangas
hlinnakan...@vmware.com wrote:
A-ha, I see. And this can happen without INSERT ON CONFLICT, too? In that
case, one of the transactions is bound to error and roll back anyway, but
you get a deadlock
On Mon, Jan 5, 2015 at 5:53 PM, Peter Geoghegan p...@heroku.com wrote:
It's not all that easy to recreate, even with my custom
instrumentation. With fsync=off, it occurs somewhat infrequently with
a custom instrumentation Postgres build:
pg@hamster:~/jjanes_upsert$ perl count_upsert.pl 8
On Thu, Jan 1, 2015 at 11:17 PM, Peter Geoghegan p...@heroku.com wrote:
The attached patch fixes Jeff's test case, as far as it goes. But, as
I mentioned, it's not intended as a real fix. For one thing, I have
made multiple changes to HeapTupleSatisfies*() routines, but haven't
fully
On Sat, Jan 3, 2015 at 10:16 PM, Peter Geoghegan p...@heroku.com wrote:
I looked at the code in more detail, and realized that there were old
bugs in the exclusion constraint related modifications. I attach a
delta patch that fixes them. This is a combined patch that is all that
is needed to
On 01/03/2015 02:43 AM, Peter Geoghegan wrote:
On Thu, Jan 1, 2015 at 11:17 PM, Peter Geoghegan p...@heroku.com wrote:
I've been working on fixing the bugs that Jeff Janes found [1] with
approach #2 to value locking [2]. Approach #1 was unaffected.
Unfortunately, exclusion constraints (which
On 01/03/2015 12:00 PM, Peter Geoghegan wrote:
Two concurrent exclusion constraints inserters can easily insert at
exactly the same time, and then wait on each other's xact, and then
deadlock. That can't happen with B-Tree inserts because the checking
and insertion happen at the same time, when
On Sat, Jan 3, 2015 at 1:29 AM, Heikki Linnakangas
hlinnakan...@vmware.com wrote:
Please explain in words of one syllable how the deadlock arises. Then,
please find a way to fix it.
I believe that the deadlock arises because there is no choke point.
Exclusion constraint insertions always insert
On Sat, Jan 3, 2015 at 2:41 AM, Heikki Linnakangas
hlinnakan...@vmware.com wrote:
A-ha, I see. And this can happen without INSERT ON CONFLICT, too? In that
case, one of the transactions is bound to error and roll back anyway, but
you get a deadlock error instead of the constraint violation
On Sat, Jan 3, 2015 at 12:42 PM, Peter Geoghegan p...@heroku.com wrote:
There are probably additional factors that make the situation more
complicated than it is for the ON CONFLICT patch, but it's clear to me
that the mutual dependency that can be involved with two ordinary
exclusion
On Thu, Jan 1, 2015 at 11:17 PM, Peter Geoghegan p...@heroku.com wrote:
I've been working on fixing the bugs that Jeff Janes found [1] with
approach #2 to value locking [2]. Approach #1 was unaffected.
Unfortunately, exclusion constraints (which only value locking
approach #2 has support for,
10 matches
Mail list logo
