Re: PL/PgSQL, Inheritance, Locks, and Deadlocks - Mailing list pgsql-general

From Thomas F.O'Connell
Subject Re: PL/PgSQL, Inheritance, Locks, and Deadlocks
Date
Msg-id f97c86cf4fcd08a40532a950ab3cbb7e@sitening.com
Whole thread Raw
In response to Re: PL/PgSQL, Inheritance, Locks, and Deadlocks  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-general
One thing that is curious, though, is that when the AccessShareLock is
acquired by the stored procedure on an unrelated linking table, there
is also an AccessShareLock acquired on the primary key of the groups
table. The latter lock is understandable, but why would the procedure
need any locks whatsoever on linking tables on which it has no direct
effect (either reading or writing)?

-tfo

--
Thomas F. O'Connell
Co-Founder, Information Architect
Sitening, LLC
http://www.sitening.com/
110 30th Avenue North, Suite 6
Nashville, TN 37203-6320
615-260-0005

On Feb 2, 2005, at 9:41 AM, Tom Lane wrote:

> "Thomas F.O'Connell" <tfo@sitening.com> writes:
>> The linking table is a pure linking table. It has a user_id and a
>> group_id, each a foreign key. The user_id ties to the appropriate
>> subclass user table. The group_id ties to the groups table, which is
>> not part of an inheritance hierarchy. A multicolumn primary key covers
>> both foreign keys in the linking table, and the secondary column of
>> the
>> key also has its own index.
>
> Inserts/updates in a table that has a foreign key result in locks on
> the
> referenced rows in the master table.  Could this explain your problem?
>
>             regards, tom lane


pgsql-general by date:

Previous
From: Tom Lane
Date:
Subject: Re: Invalid headers and xlog flush failures
Next
From: Tom Lane
Date:
Subject: Re: capturing/viewing sort_mem utilization on a per query basis