Hi Hadrian, Thamayanthy, The locking mechanism is to prevent concurrent modification of the instance state from threads operating on the same process instance from two nodes. Currently the instance lock which prevent multiple threads from concurrently operating on the same process instance is handled by InstanceLockManager class which only works for one node.
https://svn.apache.org/repos/asf/ode/branches/ode-1.3.6.x/bpel-runtime/src/main/java/org/apache/ode/bpel/engine/InstanceLockManager.java I found a reference in the riftsaw jira for this as well. https://issues.jboss.org/browse/RIFTSAW-456 Transaction rollback and retry mechanism from the scheduler should recover the instance in this scenario as noted in the jira, however, there would still be exceptions written to the log which should be preventable though an improved locking technique which works on the cluster. Regards Nandika On Thu, Mar 20, 2014 at 9:22 PM, Thamayanthy Sripalan (JIRA) < j...@apache.org> wrote: > > [ > https://issues.apache.org/jira/browse/ODE-563?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13941859#comment-13941859] > > Thamayanthy Sripalan commented on ODE-563: > ------------------------------------------ > > Hi Hadrian Zbarcea, > Thank you for your quick reply and spending your time for me in this last > hours. > > What I meant in my 3rd feature is that, if a process need to be executed > in different nodes (as you mentioned) and if it is a different thread they > cannot access the same instance at the same time, so that we need to have a > logging mechanism. > I will have a look at the cluster composition and I was thinking to handle > the nodes leaving and joining the cluster using Hazelcast and also I will > look at the zookeeper also. > > I have created a thread in the dev@ode.apache.org mailing list with the > subject > [GSoC 2014][ODE 563: Clustering for Apache ODE] > > Thank you. > > > Clustering > > ---------- > > > > Key: ODE-563 > > URL: https://issues.apache.org/jira/browse/ODE-563 > > Project: ODE > > Issue Type: Bug > > Components: Deployment > > Affects Versions: 1.2 > > Reporter: Sabir Ahamed > > Assignee: Jeff Yu > > Labels: gsoc2014 > > Fix For: 1.4 > > > > > > We need to deploy Apache ODE in clustered environment to meet > requirements for one of our projects. > > Per FAQ, the clustering support is under development, but it is not > listed in the Roadmap. > > We need to know if clustering support will be provided in the next > release i.e. 1.3. > > We will greatly appreciate if you let us know when this support will be > added or if there is any workaround. > > > > -- > This message was sent by Atlassian JIRA > (v6.2#6252) >
