Re: [infinispan-dev] Ideas for locking improvements
On 18 May 2011, at 17:23, Manik Surtani wrote:
On 18 May 2011, at 13:32, Sanne Grinovero wrote:
>> 1. Suggesting deferring local locks till prepare-time: wouldn't this create
a potentially large number of transaction failures? Since write skews and overwriting may
become a problem if this is allowed.
>
> I agree, but as far as I understood by talking to Mircea this is what
> the current implementation does: it acquired the locks locally but the
> key owners don't know about it until commit time.
> So from that I inferred that - while it surprised me - that if you're
> able to handle that then you should be able to handle the local locks
> using the same logic (defferring consistently).
True, but the way it is right now, at least in the non-clustered case transactions have a
much greater chance of completion.
No harm in 2 separate locking schemes for clustered and non-clustered though.
I don't see how this would cause more failures even on the non-clustered mode:
competing transactions fight for same locks in both approaches, just that "the
battle" is delayed to prepare time. Especially if we add the lock-reordering
optimisation that would result in no-deadlocks, I think the throughput should be better
in both local and clustered scenarios.