Re: Proposal: String key space for advisory locks

On Oct 27, 2009, at 4:37 PM, Merlin Moncure wrote:
> 'as is', advisory locks is a fantastic feature that can be used for
> signaling, mutexing, etc that are relatively difficult things to do in
> the transactional world of sql. My main gripe is that the 'shared id'
> method for doing record pessimistic locks is basically a nuclear
> missile pointed at your shared buffers if you don't have lot of
> discipline in the queries that lock IDs. Maybe this argues for more
> of a 'sql exposed' pessimistic lock feature that operates on similar
> level as 'for update'...I'm not sure...curious what thoughts you have
> about improving them.

Advisory locks have, as you say, a raft of very useful characteristics:

1. Enforced as much or as little as you wish, depending on your
application design.
2. Race-condition-free.
3. Cleaned up automatically on session end.

Of course, 2^64 potential entries are enough for anyone. The
usability issue comes when you have multiple domains that you want to
apply advisory locks to in a single database. For example, if you
have multiple tables (one of which, just for example, has a character
pk), and perhaps some inter-client mutex signaling for things that are
outside of a transactional model ("this client is importing a file
from an outside source, so don't you do it"), it's unappealing to have
to come up with ways of representing those in a 64-bit namespace.

Hashing isn't a terrible solution, assuming collisions don't become an
issue; a well-designed hashtext64() would help a lot.
--
-- Christophe Pettus
xof(at)thebuild(dot)com