Users - price rules not updating across cluste

[Loren Davie]

Hi,

I have a three-server cluster of ofbiz instances.  When a price rule  
is updated/created on the first server, its not being reflected on  
the other two servers.

Restarting the ofbiz instances on the other two servers clears the  
issue (the price rule update is reflected).  This leads me to believe  
its some sort of memory cache issue.

Any assistance would be much appreciated.  Thanks.


Loren Davie
Tech Lead
[hidden email]
212-598-2800 x206



 
_______________________________________________
Users mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/users

[Andrew Sykes]

Loren,

Caching sounds likely.

You can check whether it's a cache issue by, clearing the cache without
restarting...

webtools > cache maintenance > clear all caches.

Assuming this is 3 freestanding instances sharing a DB, you could locate
the service doing the update and create a SECA which drops a job in the
JobSandbox of all three instances to clear their caches.

Let me know if you need more detail than that...
--
Kind Regards
Andrew Sykes <[hidden email]>
Sykes Development Ltd
http://www.sykesdevelopment.com

 
_______________________________________________
Users mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/users

[Loren Davie]

Hi Andrew,

>
> You can check whether it's a cache issue by, clearing the cache  
> without
> restarting...
>
> webtools > cache maintenance > clear all caches.

Actually I recently tried that, to no effect.  So far only a server  
bounce seems to clear out whatever is being cached.

I was looking at the DistributedCacheClear feature in the  
entityengine.xml delagator element.  It sounds like it might be  
appropriate (switching the distributed-cache-clear-enabled=true, and  
friends).  It looks like it relies on JMS.

Does Ofbiz ship with a JMS provider installed?  Does this sound like  
a promising avenue to explore?

>
> Assuming this is 3 freestanding instances sharing a DB, you could  
> locate
> the service doing the update and create a SECA which drops a job in  
> the
> JobSandbox of all three instances to clear their caches.
>
> Let me know if you need more detail than that...
> --
> Kind Regards
> Andrew Sykes <[hidden email]>
> Sykes Development Ltd
> http://www.sykesdevelopment.com
>
>
> _______________________________________________
> Users mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/users
>

Loren Davie
Tech Lead
[hidden email]
212-598-2800 x206



 
_______________________________________________
Users mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/users

[David E. Jones]

If it is an entity value cache then clearing the cache manually will have an effect. Are you sure you cleared the right one, or cleared them all?

The distributed cache clear of the entity engine is what you would/should use in a multiple server environment. This takes a little effort to setup, and is most efficient for larger pools with JMS, but JMS isn't necessary as the Service Engine can use RMI to call the remote services (using a service group or whatever). This does require some knowledge of the service engine.

-David


Loren Davie wrote:

> Hi Andrew,
>
>> You can check whether it's a cache issue by, clearing the cache  
>> without
>> restarting...
>>
>> webtools > cache maintenance > clear all caches.
>
> Actually I recently tried that, to no effect.  So far only a server  
> bounce seems to clear out whatever is being cached.
>
> I was looking at the DistributedCacheClear feature in the  
> entityengine.xml delagator element.  It sounds like it might be  
> appropriate (switching the distributed-cache-clear-enabled=true, and  
> friends).  It looks like it relies on JMS.
>
> Does Ofbiz ship with a JMS provider installed?  Does this sound like  
> a promising avenue to explore?
>
>
>
>> Assuming this is 3 freestanding instances sharing a DB, you could  
>> locate
>> the service doing the update and create a SECA which drops a job in  
>> the
>> JobSandbox of all three instances to clear their caches.
>>
>> Let me know if you need more detail than that...
>> --
>> Kind Regards
>> Andrew Sykes <[hidden email]>
>> Sykes Development Ltd
>> http://www.sykesdevelopment.com
>>
>>
>> _______________________________________________
>> Users mailing list
>> [hidden email]
>> http://lists.ofbiz.org/mailman/listinfo/users
>>
>
> Loren Davie
> Tech Lead
> [hidden email]
> 212-598-2800 x206
>
>
>
>  
> _______________________________________________
> Users mailing list
> [hidden email]
> http://lists.ofbiz.org/mailman/listinfo/users

 
_______________________________________________
Users mailing list
[hidden email]
http://lists.ofbiz.org/mailman/listinfo/users