On Fri, May 25, 2012 at 1:29 PM, Mladen Turk <mt...@apache.org> wrote: > On 05/25/2012 11:28 AM, Martin Knoblauch wrote: >> >> On Fri, May 25, 2012 at 6:45 AM, Mladen Turk<mt...@apache.org> wrote: >>> >>> On 05/24/2012 09:40 PM, Anthony J. Biacco wrote: >>>> >>>> >>>> >>>> >>>> I'm still puzzled as to why this behavior just changed between .35 and >>>> .36 >>>> >>> >>> OK, but if you follow the recommended configuration >>> by making sure that workers which are members of lb are not >>> listed inside worker.list, does it works? >>> >>> >> >> The question is, is this a recommendation or a restriction that >> should be enforced? Maybe a warning about this, apparently common, >> [mis-]configuration should be issued at module configuration time? >> > > Hmm, I still didn't get the response if it works in case > member workers are not in the worker.list. >
I'm not the OP, although I believe he wrote so. > But you are right. Before 1.2.36 we just created another worker > slot in shared memory. Now it's allocated 'by name' so it can > create problems if the same name is declared both as standalone > ajp13 worker and load balancer member. > This would explain the changed behaviour. As I said, if this is a non-no with 1.2.36+, there should be a watning in the logs, or an outright refusal to load. > I'm sure that once we had 'must not be in worker.list', > but someone changed that to 'should' inside > http://tomcat.apache.org/connectors-doc/reference/workers.html > (see balance_workers directive) > > Probably needs to be changed back, or a strong warning added. Cheers Martin --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org For additional commands, e-mail: users-h...@tomcat.apache.org