When re-imaging a host pool and processing hosts in groups there is an option “Set hosts to drain mode while waiting for users to log off”. We use this option frequently, as we are not permitted to force users to log off during the process.
Unfortunately, auto-scale turns off drain mode / activates these hosts again after while. This can cause the re-imaging task to time out as new users connect to hosts with an old image. This is particularly problematic when employees work in shifts and access AVD 24/7.
Furthermore, it seems as though auto-scale prefers to disable drain mode on a host rather than start a session host (which has probably already been reimaged) to realign with its session/capacity requirements.
I would like to have an option to exclude hosts in drain mode from auto-scaling that can be applied to the host pool.
In my opinion, hosts are put in drain mode for a reason (e.g. error analysis, testing or waiting for re-imaging) and should not be activated again by auto-scale.
Comments (2 comments)