I was sitting at my desk today, waiting for (ironically enough) a client's new Skype for Business install to complete in a far far away country when I decided - hey, I am not expecting calls today, why not do a quick in-place upgrade to S4B?
Quick was not the operative word here. For reference, I have two pools - once user pool and one Persistent Chat pool (need it for demos :)), and edge pool, and a few trusted apps. Once I began the process - installing the admin tools on an admin server (not Lync), I upgraded the topology for the two pools, published, and so far so good.
The required KB2982006 was not installed on my FE servers so that was where we started, which required a reboot. Had I been wise, I would have disabled the Lync services so I would not have to wait for them to start post reboot of the server only to shut them down again so I could begin the upgrade process. I started the process on both pools, all servers, all at the same time. This was not an issue since all the services are shut down anyway, so there was no apparent communication occurring anyway.
The process started at 2:30pm my local time and the PChat pool finished approximately 30 mins later (less to uninstall and reinstall). However, the user services pool ran for two hours. It appeared that the servers were doing little to nothing during every step so I can only assume there is some fail-safe code slowing the process of uninstall and reinstall down. As a reference, the installation of the new Skype pool was completed (along with the Edge) in under one hour (granted basic install, no config, no uninstall).
I am happy to say that after the long wait, everything came up as expected and worked as expected. The edge pool was the last thing that needed an upgrade but I was waiting to get the inside pools completed prior to starting that process. I suspect it will not take long but will complete tonight and post my timings.
In short - make sure you have the requirements met for in place upgrading and the time set aside. Since the entire pool is down during the process you will have some sort of outage unless users are rehomed.
The upgrade of the Edge pool went as expected. The total time for upgrade was 30 mins and like the inside pools above, both servers were upgraded at the same time. I did notice that when I upgraded the Edge pool in the topology, the Skype-Skype Federation Search was automatically enabled. While this is a feature I do want, if you do not, or perhaps do not have the port open on the edge servers (outbound 443), then this is something you would want to disable before publishing.
Greg on May 12 commented : "Brian, is there anything new with PChat?"
Brian responded : "@Greg Unfortunately no, the PChat management and usability is the same. That means the architecture is no different as well."
Shawn Harry commented on May 20: "Yammer duplicates a lot of pChat functionality. I doubt MSFT will develop it any further in further iterations of SfB as Yammer is now their strategic offering for Enterprise internal communications/collaboration. It'll likely be de-emphasized with support remaining in place for the Large Enterprises who've invested in it until its finally deprecated."
Brian responded : "@Shawn This is true for the general masses but PChat was originally developed for the Financial Services industry and I doubt it will be going anywhere any time soon. Yammer provides little to no ethical boundaries and or security in the way PChat does and to me, provides the corporate Facebook experience in a generic/noisy way.