Hi,
I have a few dedicated assignment linked clone pools w/ persistent disks . In the past, when I've done a recompose to update the pool's base image, I've always used the 'Force users to log off' option, announced the downtime to the group of users, and did all the updates for the pool at one time. I did this so I could monitor the event in vsphere to make sure everything went well. It also didn't have a tremendous impact on the users, as I scheduled the downtime outside their regular working hours.
I have a new dedicated assignment linked clone pool for a group that operates nearly 24/7. No individual user needs to stay on for that long, obviously, but the pool has at least one user most of the day. I'd like to avoid having downtime in which everyone has to be off, if possible.
I've experimented with the "Wait for users to log off" option in the recompose menu. From what I saw, it creates the new replica and then recomposes each desktop as it becomes available without denying access to the entire pool. I left a test user logged onto one desktop in my test pool during a recompose for over 30 hours, and it didn't seem to cause a problem. After I logged off, the desktop was updated, and the old replica was deleted (as this was the last desktop to be udpated).
If I recomposed like this with this group, I would expect the recompose operation would take around 24 hours, as a few users will be working at any given point, and some others will forget to log off at the end of their shift.
Does anyone recompose like this? Does it cause any problems for you? Is there anything I should be concerned about with regard to the stretching the operation out so long?
