how to pass sessionFactory into ProcessEngine instance?
by guoyou.jiang
i use iBatis with jbpm, and create a ConnectionProvider for jbpm
Hibernate, i found when i use iBatis transaction rollback, the jbpm not
rollback, i use JDBCTransaction.
i know the spring can pass sessionFactory into jbpm ProcessEngine, but
how can i pass my own sessionFactory with no use spring?
17 years, 2 months
[JBoss jBPM] - Re: Maintenance and Repair Tools
by kukeltje
Are 1,2 and 3 valid for all DBMS you use? Would be great if you could confirm that. And maybe also what the default locking mechanisms are on specific database.
Surprises me that there is no slowdown btw. And I'm also curious what additional indexes you created that are not already on jBPM. Since I was under the impression that FK relations had an index.
Regarding the tool, finding inconsistencies is (imo) a specific problem and not a general purpose tool. Most (all?) inconsistencies (PK-FK violations, required variables not persisted etc...)should not arise if the database and transactions are configured correctly. So I'm not sure what kinds of inconsistencies you are talking about.
DBMS have tools to identify some low level issues like partially corrupt tables etc, but I get the impression that is not what you mean.
View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4249999#4249999
Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4249999
17 years, 2 months
[JBoss jBPM] - Re: Maintenance and Repair Tools
by michaelholtzman
kukeltje:
anonymous wrote : - Would be great if you could elaborate a little on the way creating the indexes impacted performance more than just the delete (thankfully it was not the coroner you had to call ;-))
(1) When we start up the jBPM java app, we load information about all the running instances from JBPM into our application. This is much faster after indexing the table (generally anything that gathers components of a process instance or process definition is significantly faster).
(2) Eliminated (almost) deadlocks on tables by enabling row locking
(3) No downgrade in run-time performance (process instances do not run slower due to overhead from indexing)
anonymous wrote : - I'm not a fan of serialzation (I've had an incompatibility once due to upgrading a jdk). If needed, I marshal the objects, or use explicit hibernate mappings
When we started this project no one really understood how hibernate worked, much less how to configure it. I think that any new variable classes we add will be implemented the same way as the native variables.
anonymous wrote : - Creating an example is not trivial, but you expect 'us' to have/know tooling for debugging it.... ;-)
Well, I wasn't looking for a tool for a specific problem, I was wondering if there were any general purpose DB tools for jBPM. (E.g., in a previous job we had a tool that would 'walk' the database and identify any inconsistencies that might cause future problems).
View the original post : http://www.jboss.org/index.html?module=bb&op=viewtopic&p=4249959#4249959
Reply to the post : http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=4249959
17 years, 2 months
