Ok to flip pg_constraint.condeferrable on 9.1? - Mailing list pgsql-hackers

From Jerry Sievers
Subject Ok to flip pg_constraint.condeferrable on 9.1?
Date
Msg-id 86eh1gn6v0.fsf@jerry.enova.com
Whole thread Raw
List pgsql-hackers
Hackers;  as per $subject...

We have an FK defined on a table large enough and 24x7  as to make a
redefining of same constraint  a painful solution.

Ran into a case where defining as deferrable initially immediate and
just running one batch job with deferred firing would solve a
concurrency problem that we discovered.

Grabbing a quick exclusive lock on the 2 related tables would not be a
problem if same might help avoid bad side-effects.

Developers are already working on a chunking solution to avoid the
long-running transaction that gave rise to this but I'd like  to
consider this if  it's not risky. 

Comments?

-- 
Jerry Sievers
Postgres DBA/Development Consulting
e: postgres.consulting@comcast.net
p: 312.241.7800



pgsql-hackers by date:

Previous
From: Andrew Dunstan
Date:
Subject: Re: json/jsonb/hstore operator precedence
Next
From: Robert Haas
Date:
Subject: Re: get_fn_expr_variadic considered harmful