[infinispan-dev] Issue with cache blocks for local read-only cache

Galder Zamarreño galder at redhat.com
Mon Sep 24 06:01:33 EDT 2012 

On Sep 21, 2012, at 3:43 PM, Sanne Grinovero <sanne at infinispan.org> wrote:

> On 20 September 2012 17:38, Andrig Miller <anmiller at redhat.com> wrote:
>> 
>> 
>> ----- Original Message -----
>>> From: "Galder Zamarreño" <galder at redhat.com>
>>> To: "Andrig Miller" <anmiller at redhat.com>
>>> Cc: "Steve Ebersole" <steve at hibernate.org>, "John O'Hara" <johara at redhat.com>, "Jeremy Whiting"
>>> <jwhiting at redhat.com>, "infinispan -Dev List" <infinispan-dev at lists.jboss.org>
>>> Sent: Thursday, September 20, 2012 6:48:59 AM
>>> Subject: Re: [infinispan-dev] Issue with cache blocks for local read-only cache
>>> 
>>> 
>>> On Sep 19, 2012, at 4:20 PM, Andrig Miller <anmiller at redhat.com>
>>> wrote:
>>> 
>>>> Yes, I can see how that can happen, if the data is deleted from
>>>> outside the application.
>>> 
>>> ^ The issue does not only happen if the data is deleted outside the
>>> application. As indicated in
>>> https://hibernate.onjira.com/browse/HHH-3817, this can happen with
>>> two competing transactions.
>>> 
>>>> If you cache something as READ_ONLY, and it gets deleted, that
>>>> doesn't fit the definition of READ_ONLY though.  You are using the
>>>> wrong cache concurrency strategy.
>>>> 
>>>> Even that issue outlines the scenario where the collection is
>>>> updated, which means its not a READ_ONLY.
>>> 
>>> I think the update is irrelevant here. The issue is related to
>>> putFromLoad + remove, which both AFAIK, are allowed in READ_ONLY
>>> (remember that we had the discussion on whether remove should be
>>> allowed in a READ_ONLY cache:
>>> https://hibernate.onjira.com/browse/HHH-7350).
>>> 
>> 
>> Yes, remove can be done, its just update that matters to READ_ONLY.  One thing I thought about was I thought we were using MVCC for this stuff.  Any transaction that reads from the cache, while something is being added/removed, should be reading the read consistent image, and should never wait on a lock, correct?  We see all the threads in our thread pool sitting in a blocked state based on this locking.

I'm not 100% sure which locking are you talking about, but if you're refering to the lock in https://dl.dropbox.com/u/30971563/specjent_block.png, that's related to the 2LC integration, not Infinispan itself.

If you're talking about threads waiting for a lock somewhere else, please provide more details.

I have some short-term ideas to improve the 2LC integration code, but I wanna check with Brian first.

Long term, I think https://issues.jboss.org/browse/ISPN-506 will be necessary to provide a lock-free solution to these edge cases in such way that 'newer' removes cannot be overridden by 'old' putFromLoad calls. However, I'm intrigued by the fact that JBoss Cache OL had the capability of being given a version externally, but the 2LC code for JBoss Cache OL still used this PutFromLoadValidator logic. Again, something I need to check with Brian.

>> 
>> That really shouldn't be happening with MVCC.
> 
> This would be correct, if only Infinispan would support MVCC, but
> that's not the case today.

Sanne, don't understand why you say that. 

In Infinispan, we make copies when writing so that existing transactions can carry on working with a read value without problems and no locking.

> 
> Sanne
> 
>> 
>> Andy
>> 
>>>> 
>>>> Andy
>>>> 
>>>> ----- Original Message -----
>>>>> From: "Galder Zamarreño" <galder at redhat.com>
>>>>> To: "infinispan -Dev List" <infinispan-dev at lists.jboss.org>
>>>>> Cc: "Steve Ebersole" <steve at hibernate.org>, "John O'Hara"
>>>>> <johara at redhat.com>, "Andrig Miller" <anmiller at redhat.com>,
>>>>> "Jeremy Whiting" <jwhiting at redhat.com>
>>>>> Sent: Wednesday, September 19, 2012 2:48:37 AM
>>>>> Subject: Re: [infinispan-dev] Issue with cache blocks for local
>>>>> read-only cache
>>>>> 
>>>>> This is code written in JBoss Cache days to deal with situations
>>>>> where putFromLoad might try to store stale data (if data is
>>>>> deleted
>>>>> in between a database read and putFromLoad being called).
>>>>> 
>>>>> Indeed this can happen with read only data, and has nothing to do
>>>>> with clustering.
>>>>> 
>>>>> The original issue is:
>>>>> https://hibernate.onjira.com/browse/HHH-3817
>>>>> 
>>>>> I'll check how this can be improved.
>>>>> 
>>>>> Cheers,
>>>>> 
>>>>> On Sep 18, 2012, at 4:55 PM, Sanne Grinovero
>>>>> <sanne at infinispan.org>
>>>>> wrote:
>>>>> 
>>>>>> There seems to be a single (global) reentrant lock which is
>>>>>> acquired
>>>>>> on any put; I don't know much about the module design, but there
>>>>>> is
>>>>>> a
>>>>>> comment close to the lock acquisition mentioning a need to flush
>>>>>> an
>>>>>> operations queue to avoid a deadlock.
>>>>>> Could it just make sure to reorder keys, so that deadlocks are
>>>>>> avoided?
>>>>>> 
>>>>>> On 18 September 2012 16:26, Manik Surtani <manik at jboss.org>
>>>>>> wrote:
>>>>>>> Agreed.
>>>>>>> 
>>>>>>> This locking appears to be in the 2LC integration code to guard
>>>>>>> a
>>>>>>> local
>>>>>>> collection though - there must be better, non-blocking ways to
>>>>>>> guard this -
>>>>>>> if it is needed at all for RO entities.
>>>>>>> 
>>>>>>> - M
>>>>>>> 
>>>>>>> On 18 Sep 2012, at 15:18, Andrig Miller <anmiller at redhat.com>
>>>>>>> wrote:
>>>>>>> 
>>>>>>> What I find interesting, is that in this use case, READ_ONLY
>>>>>>> concurrency
>>>>>>> strategy and a local cache (no invalidation and no replication),
>>>>>>> there is no
>>>>>>> need to do any locking.
>>>>>>> 
>>>>>>> If we can get to a solution without any locking that would be
>>>>>>> ideal.
>>>>>>> 
>>>>>>> Andy
>>>>>>> 
>>>>>>> ________________________________
>>>>>>> 
>>>>>>> From: "Manik Surtani" <manik at jboss.org>
>>>>>>> To: "Ståle W. Pedersen" <spederse at redhat.com>
>>>>>>> Cc: "Galder Zamarreño" <galder at redhat.com>, "John O'Hara"
>>>>>>> <johara at redhat.com>, "Jeremy Whiting" <jwhiting at redhat.com>,
>>>>>>> "Andrig Miller"
>>>>>>> <anmiller at redhat.com>, "Steve Ebersole" <steve at hibernate.org>,
>>>>>>> "infinispan-dev" <infinispan-dev at lists.jboss.org>
>>>>>>> Sent: Tuesday, September 18, 2012 8:13:43 AM
>>>>>>> Subject: Re: Issue with cache blocks for local read-only cache
>>>>>>> 
>>>>>>> 
>>>>>>> Looking at your profiler snapshot, these locks are in the
>>>>>>> Hibernate 2nd
>>>>>>> level cache implementation for Infinispan.  Galder, any ideas?
>>>>>>> 
>>>>>>> - M
>>>>>>> 
>>>>>>> On 18 Sep 2012, at 13:43, Ståle W. Pedersen
>>>>>>> <spederse at redhat.com>
>>>>>>> wrote:
>>>>>>> 
>>>>>>> hi galder and manik, sorry for sending this mail to so many, but
>>>>>>> we've ran
>>>>>>> into a issue that prevents us from further scaling of the
>>>>>>> specjenterprise2010 benchmark.
>>>>>>> 
>>>>>>> so when doing specjenterprise2010 benchmark testing we've seen a
>>>>>>> lot of
>>>>>>> blocks caused by the entity/query cache. we've been testing with
>>>>>>> only
>>>>>>> caching a simple entity bean that's read-only and queries
>>>>>>> related
>>>>>>> to this
>>>>>>> entity (selects).
>>>>>>> 
>>>>>>> here is a screenshot of the hotspot
>>>>>>> found:https://dl.dropbox.com/u/30971563/specjent_block.png
>>>>>>> 
>>>>>>> here is the standalone.xml:
>>>>>>> https://dl.dropbox.com/u/30971563/standalone-full.xml
>>>>>>> 
>>>>>>> here is the orm.xml:
>>>>>>> https://dl.dropbox.com/u/30971563/order_orm.xml
>>>>>>> 
>>>>>>> what we don't understand is why there are so many puts into the
>>>>>>> cache for an
>>>>>>> object that is marked as read-only. when we're testing without
>>>>>>> caching we do
>>>>>>> not see any blocks.
>>>>>>> 
>>>>>>> any help/ideas would be great. if anyone want a jprofiler
>>>>>>> snapshot
>>>>>>> of the
>>>>>>> run, let me know.
>>>>>>> 
>>>>>>> regards, ståle
>>>>>>> --
>>>>>>> JBoss Performance Team Lead
>>>>>>> JBoss by Red Hat
>>>>>>> 
>>>>>>> 
>>>>>>> --
>>>>>>> Manik Surtani
>>>>>>> manik at jboss.org
>>>>>>> twitter.com/maniksurtani
>>>>>>> 
>>>>>>> Platform Architect, JBoss Data Grid
>>>>>>> http://red.ht/data-grid
>>>>>>> 
>>>>>>> 
>>>>>>> --
>>>>>>> Manik Surtani
>>>>>>> manik at jboss.org
>>>>>>> twitter.com/maniksurtani
>>>>>>> 
>>>>>>> Platform Architect, JBoss Data Grid
>>>>>>> http://red.ht/data-grid
>>>>>>> 
>>>>>>> 
>>>>>>> _______________________________________________
>>>>>>> infinispan-dev mailing list
>>>>>>> infinispan-dev at lists.jboss.org
>>>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>>> 
>>>>>> _______________________________________________
>>>>>> infinispan-dev mailing list
>>>>>> infinispan-dev at lists.jboss.org
>>>>>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>>>> 
>>>>> 
>>>>> --
>>>>> Galder Zamarreño
>>>>> galder at redhat.com
>>>>> twitter.com/galderz
>>>>> 
>>>>> Project Lead, Escalante
>>>>> http://escalante.io
>>>>> 
>>>>> Engineer, Infinispan
>>>>> http://infinispan.org
>>>>> 
>>>>> 
>>> 
>>> 
>>> --
>>> Galder Zamarreño
>>> galder at redhat.com
>>> twitter.com/galderz
>>> 
>>> Project Lead, Escalante
>>> http://escalante.io
>>> 
>>> Engineer, Infinispan
>>> http://infinispan.org
>>> 
>>> 
>> 
>> _______________________________________________
>> infinispan-dev mailing list
>> infinispan-dev at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev


--
Galder Zamarreño
galder at redhat.com
twitter.com/galderz

Project Lead, Escalante
http://escalante.io

Engineer, Infinispan
http://infinispan.org

More information about the infinispan-dev mailing list