[infinispan-dev] Some docs on OGM

classic Classic list List threaded Threaded
8 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

[infinispan-dev] Some docs on OGM

Emmanuel Bernard
After being gently reminded by Manik several times, I've finally sat down and wrote a page on Hibernate OGM.
http://community.jboss.org/wiki/HibernateOGMGeneralInformations

It contains several sections:
 - how to use
 - how to contribute
 - what is supported
 - architecture

The wiki went wild, so if you edit it, copy it first (I know the text is blank on blank, go figure) in your favorite text-only editor and then paste back in a Wiki Markup element.
_______________________________________________
infinispan-dev mailing list
[hidden email]
https://lists.jboss.org/mailman/listinfo/infinispan-dev
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [infinispan-dev] Some docs on OGM

Sanne Grinovero
Hi, thank you very much, a nice reference.
In the final version you won't need the @Persister annotations right?

2011/1/6 Emmanuel Bernard <[hidden email]>:

> After being gently reminded by Manik several times, I've finally sat down and wrote a page on Hibernate OGM.
> http://community.jboss.org/wiki/HibernateOGMGeneralInformations
>
> It contains several sections:
>  - how to use
>  - how to contribute
>  - what is supported
>  - architecture
>
> The wiki went wild, so if you edit it, copy it first (I know the text is blank on blank, go figure) in your favorite text-only editor and then paste back in a Wiki Markup element.
> _______________________________________________
> infinispan-dev mailing list
> [hidden email]
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>

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

Re: [infinispan-dev] Some docs on OGM

Emmanuel Bernard
That's the goal but Core does not let us do that yet.

On 6 janv. 2011, at 11:39, Sanne Grinovero wrote:

> Hi, thank you very much, a nice reference.
> In the final version you won't need the @Persister annotations right?
>
> 2011/1/6 Emmanuel Bernard <[hidden email]>:
>> After being gently reminded by Manik several times, I've finally sat down and wrote a page on Hibernate OGM.
>> http://community.jboss.org/wiki/HibernateOGMGeneralInformations
>>
>> It contains several sections:
>>  - how to use
>>  - how to contribute
>>  - what is supported
>>  - architecture
>>
>> The wiki went wild, so if you edit it, copy it first (I know the text is blank on blank, go figure) in your favorite text-only editor and then paste back in a Wiki Markup element.
>> _______________________________________________
>> infinispan-dev mailing list
>> [hidden email]
>> https://lists.jboss.org/mailman/listinfo/infinispan-dev
>>
>
> _______________________________________________
> infinispan-dev mailing list
> [hidden email]
> https://lists.jboss.org/mailman/listinfo/infinispan-dev


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

Re: [infinispan-dev] Some docs on OGM

Manik Surtani
In reply to this post by Emmanuel Bernard
Nice.  I'll take it for a spin some time this week.  :)  Would be nice to see it in action.

On 6 Jan 2011, at 10:19, Emmanuel Bernard wrote:

> After being gently reminded by Manik several times, I've finally sat down and wrote a page on Hibernate OGM.
> http://community.jboss.org/wiki/HibernateOGMGeneralInformations
>
> It contains several sections:
> - how to use
> - how to contribute
> - what is supported
> - architecture
>
> The wiki went wild, so if you edit it, copy it first (I know the text is blank on blank, go figure) in your favorite text-only editor and then paste back in a Wiki Markup element.
> _______________________________________________
> 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
|  
Report Content as Inappropriate

Re: [infinispan-dev] Some docs on OGM

Galder Zamarreño
In reply to this post by Emmanuel Bernard
Emmanuel, good stuff on getting this done.

Looking at the code, it seems like all entities are stored within the same Infinispan Cache instance called "ENTITIES". In the second level cache we separate each entities 2LC cache into a separate cache instance? I think the latter would make it easier to support situations where certain entities are deployed in classloader isolation and some others are not, and might make it easier to define eviction/expiry settings on a per entity type.

I cloned the code last week but cannot find OgmcollectionPersister.class ??

On Jan 6, 2011, at 11:19 AM, Emmanuel Bernard wrote:

> After being gently reminded by Manik several times, I've finally sat down and wrote a page on Hibernate OGM.
> http://community.jboss.org/wiki/HibernateOGMGeneralInformations
>
> It contains several sections:
> - how to use
> - how to contribute
> - what is supported
> - architecture
>
> The wiki went wild, so if you edit it, copy it first (I know the text is blank on blank, go figure) in your favorite text-only editor and then paste back in a Wiki Markup element.
> _______________________________________________
> 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
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [infinispan-dev] Some docs on OGM

Emmanuel Bernard

On 17 janv. 2011, at 11:39, Galder Zamarreño wrote:

> Emmanuel, good stuff on getting this done.

Thanks for the feedback, see below.

>
> Looking at the code, it seems like all entities are stored within the same Infinispan Cache instance called "ENTITIES". In the second level cache we separate each entities 2LC cache into a separate cache instance? I think the latter would make it easier to support situations where certain entities are deployed in classloader isolation and some others are not, and might make it easier to define eviction/expiry settings on a per entity type.

Can you detail what's the classloader isolation issue is about? FYI, I only store "primitive types" just like a database would do. I don't store any high level type.

And for the current goal, eviction policy would be tragic, we would lose data :)
But I agree, when we will target the Grid + JPA we might need some custom eviction policy per entity type though I wonder how we will be able to address query in this case.
I'll add a todo.

>
> I cloned the code last week but cannot find OgmcollectionPersister.class ??

It's on the OGM branch
git checkout ogm
https://github.com/emmanuelbernard/hibernate-core-ogm/blob/ogm/ogm/src/main/java/org/hibernate/ogm/persister/OgmEntityPersister.java


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

Re: [infinispan-dev] Some docs on OGM

Galder Zamarreño
See below:

On Jan 17, 2011, at 2:21 PM, Emmanuel Bernard wrote:

>
> On 17 janv. 2011, at 11:39, Galder Zamarreño wrote:
>
>> Emmanuel, good stuff on getting this done.
>
> Thanks for the feedback, see below.
>
>>
>> Looking at the code, it seems like all entities are stored within the same Infinispan Cache instance called "ENTITIES". In the second level cache we separate each entities 2LC cache into a separate cache instance? I think the latter would make it easier to support situations where certain entities are deployed in classloader isolation and some others are not, and might make it easier to define eviction/expiry settings on a per entity type.
>
> Can you detail what's the classloader isolation issue is about? FYI, I only store "primitive types" just like a database would do. I don't store any high level type.

Right, I see. I'd love to see what the Infinispan cache contains when running your unit tests, is there any easy way to do that from the unit tests? I think this would be handy both from a debugging perspective.

>
> And for the current goal, eviction policy would be tragic, we would lose data :)

True, but even when memory is large, people might want to configure Infinispan with a cache store to keep memory usage under control? I mean, if your DB runs out of space, the DB data is still there, you can't put more data in it. However, if Infinspan runs out of memory, data is gone unless there's other nodes in the cluster. Use of a cache store is not maybe a good idea, but we shouldn't rule out clients doing this.

> But I agree, when we will target the Grid + JPA we might need some custom eviction policy per entity type though I wonder how we will be able to address query in this case.
> I'll add a todo.

What do you mean by addressing query? How querying will work with eviction on?

>
>>
>> I cloned the code last week but cannot find OgmcollectionPersister.class ??
>
> It's on the OGM branch
> git checkout ogm
> https://github.com/emmanuelbernard/hibernate-core-ogm/blob/ogm/ogm/src/main/java/org/hibernate/ogm/persister/OgmEntityPersister.java

Read again, I'm talking about OgmcollectionPersister.class, not OgmEntityPersister.java. OgmcollectionPersister.class comes from the wiki.

>
>
> _______________________________________________
> 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
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: [infinispan-dev] Some docs on OGM

Emmanuel Bernard

On 18 janv. 2011, at 12:08, Galder Zamarreño wrote:

> See below:
>
> On Jan 17, 2011, at 2:21 PM, Emmanuel Bernard wrote:
>
>>
>> On 17 janv. 2011, at 11:39, Galder Zamarreño wrote:
>>
>>> Emmanuel, good stuff on getting this done.
>>
>> Thanks for the feedback, see below.
>>
>>>
>>> Looking at the code, it seems like all entities are stored within the same Infinispan Cache instance called "ENTITIES". In the second level cache we separate each entities 2LC cache into a separate cache instance? I think the latter would make it easier to support situations where certain entities are deployed in classloader isolation and some others are not, and might make it easier to define eviction/expiry settings on a per entity type.
>>
>> Can you detail what's the classloader isolation issue is about? FYI, I only store "primitive types" just like a database would do. I don't store any high level type.
>
> Right, I see. I'd love to see what the Infinispan cache contains when running your unit tests, is there any easy way to do that from the unit tests? I think this would be handy both from a debugging perspective.

In debugger mode, the SessionFactory contains a SessionFActoryObserver which contains the cache. But yes, some kind of easier access to the cache would be nice.

>
>>
>> And for the current goal, eviction policy would be tragic, we would lose data :)
>
> True, but even when memory is large, people might want to configure Infinispan with a cache store to keep memory usage under control? I mean, if your DB runs out of space, the DB data is still there, you can't put more data in it. However, if Infinspan runs out of memory, data is gone unless there's other nodes in the cluster. Use of a cache store is not maybe a good idea, but we shouldn't rule out clients doing this.

Right. And you think different passivation policies / cache store per entity is better than one?
I was going for simplicity so far. also beyond entities, we have collection data (associations) and indexed data.

>
>> But I agree, when we will target the Grid + JPA we might need some custom eviction policy per entity type though I wonder how we will be able to address query in this case.
>> I'll add a todo.
>
> What do you mean by addressing query? How querying will work with eviction on?

yes but I see now that you were thinking about passivating the data no simply evicting it.

>
>>
>>>
>>> I cloned the code last week but cannot find OgmcollectionPersister.class ??
>>
>> It's on the OGM branch
>> git checkout ogm
>> https://github.com/emmanuelbernard/hibernate-core-ogm/blob/ogm/ogm/src/main/java/org/hibernate/ogm/persister/OgmEntityPersister.java
>
> Read again, I'm talking about OgmcollectionPersister.class, not OgmEntityPersister.java. OgmcollectionPersister.class comes from the wiki.

oops it's OgmCollectionPersister.class, I've fixed the wiki.
https://github.com/emmanuelbernard/hibernate-core-ogm/blob/ogm/ogm/src/main/java/org/hibernate/ogm/persister/OgmCollectionPersister.java


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