Thanks, Denise. I'll make some time next week to look harder for a solution. Currently I'm adding the host as a layout parameter. ... View more

Hey Ted, I've seen some pretty old threads here about developing training (this and Finesse Developer Training) but I'm not finding any actual training courses available on Finesse anywhere online, 3 years after the dates of these threads.  I saw a couple great videos from a recent Cisco Live! and of course there is, as Mr. Lender points out, plenty of code and writing to wade through. But learning is often far more effective when one is guided by an expert.  There have been a bazillion studies on this, and probably half as many TED talks.  Would you go climb Everest without a map or a shirpa?  Nope. (Well maybe.) Here's one link with overviews of some key points: http://changelearning.ca/get-informed/understanding-human-learning/how-humans-learn-best Also, you've based the product on a slew of technologies that not everyone is born knowing; many ContactCenter developers have been living in the corporate world (where "you only need to know what we say you need to know"); so I think it's best to include some foundational instruction in those key technologies (especially since they're key!).  You could design a 2-week course (or a 1-week accelerated version) where the first half is setting up a dev environment, OpenSocial, JavaScript, HTML, CSS, Security, and Browser Anomalies and the 2nd half is Finesse architecture, Working in the Cisco Container, Developing and Deploying Gadgets, Headless Finesse, Using the Admin API, Integration points, Troubleshooting & Admin, Testing/Best Practices, etc.  The course should all be workshop-based and allow (require?) people to test out of the first half before going to the second. Overall, the format of EACH HOUR (aka milestone) of the course should be: 0 - 5 min:           What are we faced with? 5-20 min:           What's the solution we're going to implement? 20-25 min:         Here's why it works.  A few under-the-covers notes. 25-40 min:         Making something that uses it, a demo. 40-55 min:         Workshop - now you do it. 55-60 min:         Post-mortem.  Issues, solutions, pressing questions, lead-in to next milestone. Each day should end with a 90 minute practicum/test where the student does it all again in one sprint.  It should be a hard-core learning boot camp. This *could* be done in a series of videos, like WebStorm 8: Mastering AngularJS, spy-js, Grunt, and Multi-Selection Workflows - YouTube, but 80% of the time something trivial is glossed over that prevents an individual from orienting themselves for actually using the information (e.g., there's no video from WebStorm on how to set it up to import an existing huge Eclipse project).  For that, these fora are great but maybe the video could be presented at the top of a moderated discussion for even better effect. Also, I realize that some concepts/milestones might need 2 hours, and only the greatest of instructors will be able to follow the minute-by-minute breakdown, but the formula is "What's the problem / Here's how I do it / Now you try" and THIS is the BEST way to help people. Also, providing a downloadable Finesse sandbox would be awesome, or a simulator on a VM so budgetary constraints don't hamper the ability of dedicated developers to get up to speed.  CTIOS had an incredible simulation toolset that was invaluable for development efforts that began before the environment was even up!! Chris ... View more

Apparently what happened here is a side-effect of the "Magical Cache Syndrome".  :-)  It appears as though Eclipse or the CVP plugins creates a serialized cache of any jars added to a CVP Studio Project, and there may be a test in the caching mechanism that does not update the cache if only the serialVersionUID changes.  Thus the package+class name is the same, but the runtime instance becomes incompatible.  Maybe? Changing the package name of the object, rebuilding, redeploying, fixed the issue. Then again, it may not be a builder issue.  It may also be a Tomcat session issue (it uses serialization when bouncing to retain data), though with no active callers in the test environment I'm not sure why Tomcat would even retain a session, and wouldn't the "updateApp" command clobber it? Does anyone have any detailed technical information on how this entire mechanism works?  I only like magic tricks I can fully explain... :-) Thanks, Chris ... View more

Turns out the cause of this was that someone changed the agent peripheral number at the same time that the call was arriving for the agent.  Somehow this affected the PG's ability to recache the agent data. The restart mentioned above was incomplete (not a full start of the machine, just a restart of the service), and side B was at the time stuck in a "configuring" state, so there was no failover.  CallManager's CTI Services and both sides A and B had to be restarted for the old agent to be released and the new one rebuilt. This, anyway, is as close to an explanation as I've been able to come.  The full power cycle of the PGs coupled with restarting CM's CTI services successfully released the invalid agent configuration and has allowed the agent to get back to work again. Thanks, Chris ... View more

Thanks Geoff! I haven't been able to get to the router log viewer yet - hopefully it will still contain something at this late date. However, in order to start "with a clean slate", our server admin rebooted the PG pair in question and now the agent is unable to log in.  We login with LoginName and then in the OnSetAgentMode event, the Agent object returned has an agent ID that does not exist in ICM, either in current or deleted objects!  Due to this, she is unable to get past the login. I'm going to delete and recreate her agent in hopes the PG is refreshed with the correct ID for the login name.  I am not sure how to explain this to the mangerial types, where the bogus agent ID came from.  Something tells me this invalid set up was also why the call couldn't be set up, but I'm not sure why, of all the agents that experienced the issue only hers has proceeded to the invalid ID stage (others can login fine). Thanks, Chris ... View more