[infinispan-dev] Competing for cluster wide lock between state transfer and rehashing

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

[infinispan-dev] Competing for cluster wide lock between state transfer and rehashing

Galder Zamarreno
Hi,

Re: https://issues.jboss.org/browse/JBPAPP-6052

I suspect here that the Hot Rod topology cache state transfer is competing with, most likely, the rehashing of distributed caches for the cluster wide distributed lock.

Can someone (Manik?) remember why is this needed if clearly these two caches are unrelated?

Cheers,
--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache


_______________________________________________
infinispan-dev mailing list
[hidden email]
https://lists.jboss.org/mailman/listinfo/infinispan-dev
Reply | Threaded
Open this post in threaded view
|

Re: [infinispan-dev] Competing for cluster wide lock between state transfer and rehashing

Manik Surtani
Yeah the processing lock in this case is on the Transport, which is shared by multiple caches.  The lock is used to maintain a coherent transaction log during state transfer or rehash.  Which means >1 cache cannot perform a rehash or state transfer at the same time.  We may be able to loosen this constraint, but this needs careful thought.  Create a JIRA for it?

In the meanwhile just make sure your state transfer or rehash timeouts take this into consideration.

On 10 Mar 2011, at 10:48, Galder Zamarreño wrote:

> Hi,
>
> Re: https://issues.jboss.org/browse/JBPAPP-6052
>
> I suspect here that the Hot Rod topology cache state transfer is competing with, most likely, the rehashing of distributed caches for the cluster wide distributed lock.
>
> Can someone (Manik?) remember why is this needed if clearly these two caches are unrelated?
>
> Cheers,
> --
> Galder Zamarreño
> Sr. Software Engineer
> Infinispan, JBoss Cache
>
>
> _______________________________________________
> infinispan-dev mailing list
> [hidden email]
> https://lists.jboss.org/mailman/listinfo/infinispan-dev

--
Manik Surtani
[hidden email]
twitter.com/maniksurtani

Lead, Infinispan
http://www.infinispan.org




_______________________________________________
infinispan-dev mailing list
[hidden email]
https://lists.jboss.org/mailman/listinfo/infinispan-dev
Reply | Threaded
Open this post in threaded view
|

Re: [infinispan-dev] Competing for cluster wide lock between state transfer and rehashing

Galder Zamarreno
Right, I understand now. Definitely something to look into and see if it can be optimised... something for 5.x

Btw, your explanation would be a great addition to the distributedSyncTimeout config setting, people have asked it in the past - http://community.jboss.org/message/546285#546285 - I'll add it asap to the code.

On Mar 10, 2011, at 2:47 PM, Manik Surtani wrote:

> Yeah the processing lock in this case is on the Transport, which is shared by multiple caches.  The lock is used to maintain a coherent transaction log during state transfer or rehash.  Which means >1 cache cannot perform a rehash or state transfer at the same time.  We may be able to loosen this constraint, but this needs careful thought.  Create a JIRA for it?
>
> In the meanwhile just make sure your state transfer or rehash timeouts take this into consideration.
>
> On 10 Mar 2011, at 10:48, Galder Zamarreño wrote:
>
>> Hi,
>>
>> Re: https://issues.jboss.org/browse/JBPAPP-6052
>>
>> I suspect here that the Hot Rod topology cache state transfer is competing with, most likely, the rehashing of distributed caches for the cluster wide distributed lock.
>>
>> Can someone (Manik?) remember why is this needed if clearly these two caches are unrelated?
>>
>> Cheers,
>> --
>> Galder Zamarreño
>> Sr. Software Engineer
>> Infinispan, JBoss Cache
>>
>>
>> _______________________________________________
>> infinispan-dev mailing list
>> [hidden email]
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>
> --
> Manik Surtani
> [hidden email]
> twitter.com/maniksurtani
>
> Lead, Infinispan
> http://www.infinispan.org
>
>
>
>
> _______________________________________________
> infinispan-dev mailing list
> [hidden email]
> https://lists.jboss.org/mailman/listinfo/infinispan-dev

--
Galder Zamarreño
Sr. Software Engineer
Infinispan, JBoss Cache


_______________________________________________
infinispan-dev mailing list
[hidden email]
https://lists.jboss.org/mailman/listinfo/infinispan-dev