multiple instances of atleap using the same db

[Jérémy Lal]

Hi,
i'm wondering if it's manageable to use multiple instances of atleap (on different servers),
that use the same database.
Of course that would be for load-balancing...

Cheers,
Jérémy Lal


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Roman Puchkovskiy-2]

I think it would be problematic. We use Hibernate generator 'increment'
often, and here's what is written about it in Hibernate docs
(http://www.hibernate.org/hib_docs/reference/en/html/mapping.html):

<<
increment

    generates identifiers of type long, short or int that are unique
only when no other process is inserting data into the same table. Do not
use in a cluster.
>>

Your case resembles cluster case, seems to me. In any case, you will
probably have constant failures because of non-unique PKs when inserting
intensively.

Another problem is Hibernate cache which should be aware of the fact
that someone else accesses the DB. If you just run N copies of Atleap,
you will have N copies of Hibernate along with its cache.

Jérémy Lal writes:

> Hi,
> i'm wondering if it's manageable to use multiple instances of atleap (on
> different servers),
> that use the same database.
> Of course that would be for load-balancing...
>
> Cheers,
> Jérémy Lal
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@...
> For additional commands, e-mail: users-help@...
>
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...

[Jérémy Lal]

A - would there be some hibernate.transaction.factory_class that would allow for a proper locking of transactions
that allows increment to work correctly ?
i'm searching stuff on this, but i'm not very familiar with the concept.
i guess increment is difficult to replace with another cluster-safe strategy ?

B
1) make use of oscache 2 (instead of hibernate's builtin oscache provider)
http://www.opensymphony.com/oscache/wiki/Hibernate.html
2) configure oscache for clustering
http://www.opensymphony.com/oscache/wiki/Clustering.html

Thanks for your time.


On 03.02.2009 15:29, Roman Puchkovskiy wrote:

> I think it would be problematic. We use Hibernate generator 'increment'
> often, and here's what is written about it in Hibernate docs
> (http://www.hibernate.org/hib_docs/reference/en/html/mapping.html):
>
> <<
> increment
>
>      generates identifiers of type long, short or int that are unique
> only when no other process is inserting data into the same table. Do not
> use in a cluster.
>
> Your case resembles cluster case, seems to me. In any case, you will
> probably have constant failures because of non-unique PKs when inserting
> intensively.
>
> Another problem is Hibernate cache which should be aware of the fact
> that someone else accesses the DB. If you just run N copies of Atleap,
> you will have N copies of Hibernate along with its cache.
>
> Jérémy Lal writes:
>> Hi,
>> i'm wondering if it's manageable to use multiple instances of atleap (on
>> different servers),
>> that use the same database.
>> Of course that would be for load-balancing...
>>
>> Cheers,
>> Jérémy Lal
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@...
>> For additional commands, e-mail: users-help@...
>>
>>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe@...
> For additional commands, e-mail: users-help@...
>

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@...
For additional commands, e-mail: users-help@...