WebSphere Process Server: Messages sitting on queues unprocessed

We’re battling another runtime problem. We’re on 6.0.2.3 and WAS 6.0.2.23 and have a situation where our modules that enable Event Sequencing seem to deadlock when trying to pick messages up off the queue. The transactions then time-out in a sea of exceptions. We have a messaging engine cluster (set to 1 of N) and a two-node SCA cluster (M of N). It seems like the SCA cluster just can’t do anything on the messaging engine.

We’re working with IBM Support on this one. I know that there are a ton of APARs about SIBus contained in WAS 6.0.2.25 and WAS 6.0.2.27 that seem like they MAY fix the issue, but jurisprudence won’t allow us to just upgrade without assured resolution.

Author: dan

Comments

  1. We had an issue where we were using event sequencing in an environment which contained 2 WPS nodes, each running on a local ME. The event sequencing modules were only deployed on one server. Depending on the order of server startup, the event sequencing either worked or did not work.

    If the server which did not host the event sequencing modules was started first, ES failed to run because it started up on the wrong server.

    Also if the server which ES modules ran on was restarted, leaving the other one running, ES failed over to the other server and could not run because it could not see any destinations.

    When the server hosting the ES modules was restarted, the non ES server also had to be restarted so that the event sequencing failed back to the correct server. There is a fix for this now.

Leave a Reply

Your email address will not be published. Required fields are marked *