Vivek Agarwal’s Portal/Java Blog

An IBM Gold Consultant’s weblog about IBM, Lotus, WebSphere, J2EE, IT Processes, and other IT technologies

Battling WebSphere Portal

Posted by Vivek Agarwal on October 26, 2005

Just a quick update to say that we continue to battle WebSphere Portal v5.1 and making it work in a dual-master site operation. Have been busy trying to resolve one issue after another and have had no time to write more. Thought I would just vent a little bit on WPS.

We are trying to make two WPS v5.1 clusters work as replicas of each other with each site allowing updates simulatenously – i.e. dual master site operation. Our plan is to get this working perfectly in our staging environment, take file system/database backups, and then restore in production for our final deployment. It has been a challenge getting this to work and I will write more about some of the challenges later. However, for the last few weeks, we have just been having a devil of a time battling WPS bugs/limitations, and things stopping to work randomly in our staging environment.

  • Ran into issues with getting the WMM files modified in the cluster. Followed the instructions from IBM but for some reason, the updates would not show up on the clone node. Tinkered around with that for quite a while and then the updates did show up on the clone node, but we are unsure what the original issue was and what made it work. Oh well!
  • The people finder portlet stopped working after working like a charm for a couple of months. We had customized the portlet configuration and also hacked some of the jsps, and all of it had been working fine. Then sometime last week, the people finder portlet stopped working. I am not sure what happened other than us applying fixpack 2. Anyways, ended up deleting the custom configuration and reconfiguring the people finder – that seems to have fixed it. Again, oh well!
  • We are still chasing IBM to acknowledge that it is unacceptable for WebSphere Portal to unilaterally execute the home page on a 404 error (my previous blog entry talks about this one) without allowing the Portal administrator to configure the Portal behavior on a 404 error. On thinking over this one further, I believe that IBM Support’s response that it is the developer’s responsibility to fix all URL references is utterly rubbish. The Portal outputs no useful information to the developer that a bad URL even occurred – so even if you have a bad link, you will not know about it (this is not an error that the web server knows anything about either). And to top it off, there is no Portal documentation that mentions this gotcha (except my blog entry)!
  • Getting the Portal Document Manager to work in the dual master sites is proving to be a nightmare. At this point, I am beginning to wonder whether it is even feasible. The PDM database is a total mess – many tables do not have primary keys and there are many aspects that make it difficult to do peer-to-peer Oracle database replication with the PDM schema. That is the most sticky outstanding issue at this point on this project and all our attention is focused on resolving this issue. I will definitely go deeper into this issue in a future blog.

I shall be back with more soon and hopefully will have resolved the nasty PDM issue!


Sorry, the comment form is closed at this time.

%d bloggers like this: