can i have distributed caching with local locking ?

[sudheerk84]

i have usecases werein i need to data to be replicated to multiple nodes, but i am sure that only only node is writing into that caching.

So is there a configuration in infinispan to set , so that only local locks will be acquired.  Sicne we are using many nodes and JTA transactiosn , this will be a very good performance boost for our system.

[sudheerk84]

can i use this eagerLockSingleNode ?

i was checking teh documentation - http://docs.jboss.org/infinispan/5.0/apidocs/config.html

It has same description for eagerLockSingleNode and useEagerLocking .

For my use case,  setting both flags true will server the purpose ?

Editing :

I found more info teh same here

https://docs.jboss.org/author/display/ISPN/Locking+and+Concurrency#LockingandConcurrency-Explicitandimplicitdistributedeagerlocking

thanks

[galder.zamarreno]

Hmmm, I don't see how these options are really gonna help. If you can assure that only one thread will modify a cache entry at one point, you could use SKIP_LOCKING flag, but there's danger in doing that.

You should also consider the unsafe options: https://docs.jboss.org/author/pages/viewpage.action?pageId=8093877

Finally, you should check the per-invocation flags that can help optimise your cache usage: https://docs.jboss.org/author/x/IgY5

[sudheerk84]

I have multiple thread writing into teh cache locally , but no other node is writing into it.

When i use a cache.put() - I see that local lock is taken, and when i try to commit , it takes remote lock on all the other nodes.  My problem here is , i do not want it to take remote locks on any other node(since non other node updates this cache).

So i assume by setting this eagerLockSingleNode, it will take only this local lock. meanwhile any other thread on this vm trying to update will wait. Isnt that right ?