I have put the long running job in the offending component into it's own back ground thread. I will report back to see if that resolves the problem..
On Tue, Jan 28, 2014 at 2:06 PM, David Jencks <[email protected]>wrote: > I think so... registering component 1 results in component 2 becoming > satisfied and activating, calling the long-lasting activate method (more > than 5 seconds) so the service registration doesn't return. > Meanwhile another thread registers a service that could cause activation > of component 1 (even though its already in the process of activating) so > also tries to activate it. > > The thread dump should show if this is happening. > > thanks > david jencks > > On Jan 28, 2014, at 10:48 AM, Dave Smith <[email protected]> wrote: > > > If I was doing some long running method in an activate method of a > service > > that I depend on would that cause the problem as well? It is intermittent > > and seems to depend on the speed of the persons computer and network > > connection speed. > > > > > > On Tue, Jan 28, 2014 at 1:27 PM, David Jencks <[email protected] > >wrote: > > > >> That's from DS. If you turn up logging to debug level you should see a > >> thread dump in your log that hopefully will explain what's going on. > >> > >> Usually when I get this it means that a service registration has gotten > >> stuck is some code that uses the registration and then blocks on > something. > >> > >> If you need to you can increase the timeout from the default 5 seconds. > >> > >> You can configure these in framework properties or in config admin. > >> > >> ds.loglevel=debug > >> ds.lock.timeout.milliseconds=5000 > >> > >> hope this helps > >> david jencks > >> > >> On Jan 28, 2014, at 4:25 AM, Dave Smith <[email protected]> wrote: > >> > >>> I occasionally get the following message using the new SCR > >>> > >>> [<Class Name>(<service pid id))] Timeout waiting for reg change to > >> complete > >>> registered > >>> > >>> IS that coming from SCR or is that buried in equinox? > >> > >> > >> --------------------------------------------------------------------- > >> To unsubscribe, e-mail: [email protected] > >> For additional commands, e-mail: [email protected] > >> > >> > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
