ode-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Karolis Petrauskas <k.petraus...@gmail.com>
Subject Re: High availability
Date Tue, 01 Feb 2011 23:52:31 GMT
Hi Alexey,

    Thank you for your comments. Unfortunately my case does not fit to
these restrictions. I have long running process (up to 5 days)
involving human tasks, correlations, etc.

Karolis Petrauskas

On Fri, Jan 14, 2011 at 8:19 PM, Alexey Ousov <aousov@gmail.com> wrote:
> Hello Karolis.
> It can be done, but with many restrictions. All processes should be
> in-memory, all ODE instances should use local database (in-process firefox
> for example), all processes should be synchronically deployed on each ODE
> instance. In such configuration you can run only simple processes with
> single entry point, without correlation, database variables and other
> benefits. You can interact with separate external database, which can be one
> for all processes, but it should be done via external service, which
> processes can call. In front of all your ODE instances you can place Apache
> Synapse with load balanced configuration.
> Regards,
> Alexey
>> Hi all,
>>    As I understand from the ODE FAQ and mailing list archives,
>> clustering in Apache ODE is not supported currently. If this is the
>> case, I want to ask you all, if there are some ways to achieve high
>> availability without real cluster implementation? Maybe there are some
>> practices and/or techniques to achieve this by putting for example JMS
>> with Camel routing in front of several ODE instances, or are there
>> some different ways? Any hints are very appreciated.
>> Thank you in advance,
>> Karolis Petrauskas

View raw message