[infinispan-dev] conflicts resolution in DeltaAware

[Mircea Markus]

On 8 Apr 2011, at 18:07, Sanne Grinovero wrote:

> Hi Mircea,
> I remember you recently mentioned that you have been looking into ways
> to give the ability to the application to resolve updating conflicts.
> I don't think you where referring to AtomicMap or any other DeltaAware
> specifically, but it seems now that we urgently need something like
> this in OGM.
yes, what I was looking for is a way to efficiently compare the state of two partitions after a split brain. Is this your scenario?
> 
> I'm looking into the Delta interface, which defines this single method:
> 
> DeltaAware merge(DeltaAware d);
> 
> This looks great to merge an update onto an existing value; But of
> course the world is made of parallel executing updates, especially so
> when transactions are involved, so we would like to be able to deal
> with multiple concurrent updates being applied on the same read value.
> 
> I'm wondering if we could get something like
> 
> DeltaAware merge(DeltaAware originalValue, DeltaAware updateA,
> DeltaAware updateB) throws UnsolvableConflict;
hmm, Delta.merge(DeltaAware) is only invoked with lock held on the corresponding. Multiple tx won't be able to own that lock at the same time. I might be wrong about this, but with transactions, I only see this to be useful if multiple threads operate on the same key within the same transaction (this is something we don't support atm). 
> 
> As Emmanuel pointed out while implementing OGM, if two different
> updates are being applied to an AtomicMap, it's possible that the
> updates relate to different keys in the AtomicMap, and for the OGM
> usecase we would be totally fine with that and would *not* like the
> transaction to be rolled back.
not sure it is atomic map you need then, as w access to AM is serialized by acquiring lock on its key. If you want finer grained, per key access, you might want to use a Cache and not an AtomicMap.  
> From my understanding of AtomicMap, if
> two different keys are changed one transaction will fail (correct?).
the calls are serialized and applied one at a time,(this is enforced by the locking) both should be applied successfully. 
> Of course it's totally possible that both updates where going to
> affect the same keys, and in this case we want to have the operation
> rolled back.
> 
> I don't mean the API I wrote literally, I wrote it like that just to
> give a general picture of what we need; I'm not really liking the idea
> of throwing an exception on a potentially frequently occurring event,
> and also while you could chain such invocations in case of multiple
> updates arriving on the same key, I think it would be far better to be
> able to try different combinations - or let the application try to be
> clever - to try resolving as much non conflicting DeltaAware updates
> as possible.
> 
> Considering that each transaction might change multiple keys, it would
> be awesome to describe the API in such a way that either Infinispan or
> the application can be "smart" and be able to estimate which
> transaction could be removed (rolled back) to provide the minimum
> amount of rolled back transactions. I'm not even sure the policy
> should favour the maximum amount of transactions, maybe the
> application wants to favour the highest possible changeset, or a
> specific important one.
> 
> Even if we start just implementing the method signature I proposed,
> that could bring a hell of scalability improvements; possibly not only
> to HibernateOGM but as far as I understood the TreeCache as well.
> 
> Ideas?
this optimisation might be doable: if a tx1 tries to write to a ATM that is locked by tx2, if keys in tx1 do not overlap keys in tx2 then we can allow them both write access. If there's an overlap serialise the access (current behaviour). This is very similar to what the Cache's semantics though, and you might want to consider weather you want a ATM here or a Cache.
> 
> Emmanuel, do you agree this would fit the HibernateOGM needs?
> 
> Cheers,
> Sanne