How do Windows Server sysadmins handle updates that require reboots?

[KaoDome]

Hi, I think it's well known there are several sysadmins among the members of Neowin and I have always been curious about this so I decided to ask away. In a desktop/workstation setting rebooting the machine because of an update isn't that big of a deal, but server uptime is important enough that some companies sign SLAs guaranteeing a certain uptime.

 

I know (some?) Linux systems are able to patch the kernel live and can work as usual without a full system reboot for a long period of time, but what about Windows? Just like their desktop counterparts I'm sure the server line has updates that require a reboot.

 

If anybody has first or third hand experience on how it's handled I'd appreciate your input, but thanks anyway 😉

[+mram Subscriber²]

Maintenance windows are your friend.  There are many reasons for them, and SLAs are usually outside of that scope.  After that it's not so much about the machine, but about the service.  Any mission critical service running on a single box is always going to have "downtime" of some sort, but a load-balanced bank of servers where you can restart in a sequence effectively has zero downtime of the service, even if the machine itself goes down.  That's just an example.

 

This has nothing expressly to do with Windows bear in mind -- this is pretty much how datacenters are run.  There are many people who would immediately jump to how Windows servers are "evil" but patching of any kind of system is and always should be part of the management and maintenance of systems.

[Shiranui]

For our single file server at work, I just schedule a reboot at 5am, Monday morning.