Microsoft are currently distributing the following email template:
Performance, security, and quality are always top priorities for us. I am reaching out to give you an advanced notice about an upcoming planned maintenance of the Azure host OS. The vast majority of updates are performed without impacting VMs running on Azure, but for this specific update, a clean reboot of your VMs may be necessary. The VMs associated with your Azure subscription may be scheduled to be rebooted as part of the next Azure host maintenance event starting January 9th, 2018. The best way to receive notifications of the time your VM will undergo maintenance is to setup Scheduled Events <https://docs.microsoft.com/en-us/azure/virtual-machines/windows/scheduled-events> .
If your VMs are maintained, they will experience a clean reboot and will be unavailable while the updates are applied to the underlying host. This is usually completed within a few minutes. For any VM in an availability set or a VM scale set, Azure will update the VMs one update domain at a time to limit the impact to your environments. Additionally, operating system and data disks as well as the temporary disk on your VM will be retained (Aidan: the VM stays on the host) during this maintenance.
Between January 2nd and 9th 2018, you will be able to proactively initiate the maintenance to control the exact time of impact on some of your VMs. Choosing this option will result in the loss of your temporary disk (Aidan: The VM redeploys to another host and gets a new temporary disk). You may not be able to proactively initiate maintenance on some VMs, but they could still be subject to scheduled maintenance from January 9th 2018. The best way to receive notifications of the time your VM will undergo maintenance is to setup Scheduled Events <https://docs.microsoft.com/en-us/azure/virtual-machines/windows/scheduled-events> .
I have put together a list of resources that should be useful to you.
* Planned maintenance how-to guide and FAQs for Windows <https://docs.microsoft.com/en-us/azure/virtual-machines/windows/maintenance-notifications> or Linux <https://docs.microsoft.com/en-us/azure/virtual-machines/linux/maintenance-notifications> VMs.
* Information about types of maintenance <https://docs.microsoft.com/en-us/azure/virtual-machines/windows/maintenance-and-updates> performed on VMs.
* Discussion topics for maintenance on the Azure Virtual Machines forums.
I am committed to helping you through this process, please do reach out if I can be of any assistance.
Regards
<Insert signature>
In short, a deployment will start on Jan 9th that will introduce some downtime to services that are not in valid availability sets. If you are running VMs that might be affected, you can use the new Planned Maintenance feature between Jan 2-9 to move your VMs to previously updated hosts at a time of your choosing. There will be downtime for the Redploy action, but it happens at a time of your choosing, and not Microsoft’s.
For you cloud noobs that want to know “what time on Jan 9th the updates will happen?”, imagine this. You have a server farm that has north of 1,000,000 physical hosts. Do you think you’ll patch them all at 3am? Instead, Microsoft will be starting the deployment, one update domain (group of hosts in a compute cluster) at a time, from Jan 9th.
And what about the promise that In-Place Migration would keep downtime to approx 30 seconds. Back when the “warm reboot” feature was announced, Microsoft said that some updates would require more downtime. I guess the Jan 9th update is one of the exceptions.
My advice: follow the advice in the mail template, and do planned maintenance when you can.
Want to Learn About In-Place Migration, Availability Sets, Update & Fault Domains?
If you found this information useful, then imagine what 2 days of training might offer you. I’m delivering a 2-day course in Amsterdam on April 19-20, teaching newbies and experienced Azure admins about Azure Infrastructure. There’ll be lots of in-depth information, covering the foundations, best practices, troubleshooting, and advanced configurations. You can learn more here.
Actual Maintenance will start on January 10th 00:00:00 and will complete before January 11th 00:00:00. A second batch will run from January 11th until January 12th. And the last batch of 1 day will run from January 12th until January 13th.
Other VMs might be hit during a bigger window: January 15th until January 21st.
When you as customer want to influence Maintenance, take actions as described BEFORE January 9th. But be aware, not all VMs can be scheduled by the customer.