Okay, after a little bit of investigation I think I got what caused the crashes on my machine: - I removed "cluster.registerManager(manager);" from the code base because it was commented out in the original code base - I commented out "<property name="mcast-bind-addr" value="192.168.2.1"/>" from ofbiz-component.xml because it was conflicting with my local LAN setup.
After doing the above and testing the cluster, everything worked. Bug again, given that I'm treading in new territory, appreciate all the help. The new patch is available at https://issues.apache.org/jira/browse/OFBIZ-9392 Cheers, Taher Alkhateeb On Tue, Jun 13, 2017 at 9:17 AM, Taher Alkhateeb <slidingfilame...@gmail.com > wrote: > So the good thing is that the refactoring exercise surfaced this issue. > Based on feedback from Scott and Michael I suspect that the issue might be > perhaps minor such as a missing configuration or dependency that's causing > all these errors. > > I'll try to dig out the root cause and your help would be great to try and > zoom in on the problem preferably on the refactored code, but even finding > the root cause on existing code would greatly help. > > Cheers, > > Taher > > On Jun 13, 2017 4:16 AM, "Scott Gray" <scott.g...@hotwaxsystems.com> > wrote: > >> The project I'm currently on is actively using clustering (session >> replication) and I know for certain this isn't the only project nor are we >> the only vendor to use it. I'm using an older version of OFBiz though so >> I >> can't confirm if the latest codebase has it working. It can be tricky to >> set up and does have a tendency to crash ofbiz during startup if you don't >> have it configured correctly. >> >> Definitely a -1 for removal. >> >> Regards >> Scott >> >> On 13 June 2017 at 08:57, Taher Alkhateeb <slidingfilame...@gmail.com> >> wrote: >> >> > Hi Everyone, >> > >> > In reference to [1] and the refactoring work, and after multiple >> different >> > tests, I came to realize that Tomcat clustering implementation is broken >> > (both before and after the refactoring). The reasons are complex but >> they >> > essentially have to do with setting up the right parameters for the >> > cluster. So I suspect no one is using this feature? >> > >> > It is therefore, my recommendation to remove clustering from >> > CatalinaContainer which would substantially reduce complexity of both >> the >> > component configuration file and the accompanying source code which >> would >> > shave off at least 120 lines of code. >> > >> > WDYT >> > >> > [1] https://issues.apache.org/jira/browse/OFBIZ-9392 >> > >> > On Thu, Jun 8, 2017 at 3:38 PM, Taher Alkhateeb < >> > slidingfilame...@gmail.com> >> > wrote: >> > >> > > Hey folks, >> > > >> > > It was very painful and slow, but I finally got a working almost >> > > full-rewrite of the CatalinaContainer. I need help with testing, >> reviews, >> > > and code improvements. >> > > >> > > Details are found in [1]. There are no functional changes, just a >> > rewrite. >> > > I appreciate all the help and feedback. >> > > >> > > [1] https://issues.apache.org/jira/browse/OFBIZ-9392 >> > > >> > > Cheers, >> > > >> > > Taher Alkhateeb >> > > >> > > >> > > >> > >> >
