3 ms·
Ask HN: How do you lower the impact of Windows reboots on servers after updates?
A customer of mine has to run a program on a Windows server. We start the server in an OpenStack cloud, start the program, call its API from remote, get the results of the computation, destroy the server without even stopping it. Every few months we rebuild the image after spending half a day applying Windows updates. We have to start using an always on Windows server now (due to changes in licensing, no alternatives.) We are worried about the unavailability of the server during the reboots Windows does during the updates. Reboots would also break any running computation. We're all using Linux or Macs so our Windows administration skills are pretty low. Is there any standard way to defer the reboots of Windows to when we want them to happen without interfering with the update process? As a comparison, Linux can keep downloading new kernels and system libraries for months and start using the most recent ones when we finally reboot it. All it takes is a single few seconds long reboot.