Introduction Types of planned maintenance The majority of

Introduction: Types of planned maintenance The majority of updates have no impact to hosted VMs. However, there are cases where updates to the hosting infrastructure result in interference to running VMs: • VM preserving maintenance (aka PHU- Preserving Host Updates) • The virtual machine is placed into a “paused” state • RAM, Network connection, open files remain open • Virtual machine is then resumed within 30 seconds • VM restarting maintenance • VM Reboot – VM is restarted. Memory, open files, network connections are lost • VM Redeploy – VM is moved to another host. Ephemeral drive is lost as well

Planned Maintenance: Goals Better communication More Control • • Longer Notification time. Per VM status Configure recipients (not just admin) Multiple Channels • • User initiated maintenance • 14 -30 days control window • In-VM Scheduled Events Notification Email/SMS/Web. Hook Programmatically Azure Portal Microsoft Confidential Reduced impact • A single VM Restarting maintenance per year • Tighter maintenance window • Keep ephemeral disks if possible • Innovate in order to reduce impact (Live. Migration)

The planned maintenance cycle • Maintenance Cycle contains multiple iterations (waves) • Each iterations has separate scheduled and scope. • Safe Deployment • Initial iterations have smaller scope: • EUAP Regions ( Canary) First Region (PILOT) Broad rollout • Region Pairs • Enable customers to have BC/DR schemas across paired regions (within a geo) • An iteration will impact only one region in the pair • See the list of Azure region pairs. • UD Safe • Cloud Services, Availability Set and VM Scale Set all use Update Domains (UDs) • Only a single update domain is impacted at any given time

The planned maintenance iteration (wave) T 0 • Pre-emptive Window (30 days) • Starts with a notification to the user • • One notification per subscription Sent to subscription’s admin Customers can add recipients Customer can add more channels (SMS, web hook) • Maintenance window is discoverable • Discover which VMs are going to be impacted in this wave • Azure maintenance dashboard • Azure VM list blade, VM Details blade, API, Power. Shell, Cli • Enable customers to start maintenance on their VMs T 1 T 2 T 3 • Scheduled Maintenance Window • Maintenance window is discoverable • In-VM Scheduled Events raised 15 minutes prior to the actual impact.

Should I start maintenance myself ? • Initiate maintenance when: • • • You need to communicate an exact maintenance window You need to control/orchestrate the impact You need more than 30 minutes between each two VMs You have a large state stored in a local (ephemeral) disk You are running a single instance VM for a production workload • Let Azure initiate maintenance when • You have a large deployment (Availability Set, VMSS) where a single VM going down does not affect the overall availability • You are using scheduled events to proactively drain or failover VMs. Impact: User Initiated: The VM will move to another node (ephemeral drive is lost) Platform Initiate: The VM will be rebooted (ephemeral drive is kept) Note: Single VM per node (D 15, DS 15, GS 5, L 32) will be maintained in-place.

Scheduled Events: Surfacing Upcoming Events Inside a VM • Surface upcoming maintenance events from within the VM to improve Availability • A local endpoint with a simple REST API • Visibility to upcoming event across all instances: VMs, cloud service / Availability Set/ VMSS • A Not. Before time (10 -15 minutes notification) • Acknowledge completion to expedite • Potential use cases: • Graceful shutdown – save state, drain node, suspend jobs • Proactive failover – fasted failover (skip detection) • Adjust thresholds – Avoid failover in the case of VMpreserving maintenance • Cover all maintenance scenarios • Platform initiated (e. g. Host OS rollout) • Interactive calls (e. g. restart a VM) • Predictable hardware failures curl -H Metadata: true http: //169. 254/metadata/scheduledevents? api-version=2017 -03 -01 { "Document. Incarnation": {Incarnation. ID}, "Events": [ { "Event. Id": {event. ID}, "Event. Type": "Reboot" | "Redeploy" | "Freeze", "Resource. Type": "Virtual. Machine", "Resources": [{resource. Name}], "Event. Status": "Scheduled" | "Started", "Not. Before": {time. In. UTC}, } ] }

Learn more • Azure Friday – Planned Maintenance (Video) • Planned Maintenance concepts • Planned Maintenance – How-To for Windows and Linux • Planned Maintenance – Blog post

The s lide w ill be This s repla lide is req ced o u nsite ired. Do N throu gh Si OT delete lver F ox Pr or alter th oduc tions e slide. with an up da http: //myignite. microsoft. com https: //aka. ms/ignite. mobileapp ted Q R cod e.

Set activity log alert with Type=Maintenance and Status = ALL Add recipients and channels to your alert

New ‘Planned Maintenance Dashboard’

See Impacted resources

Alternatively, use the VM List view

Maintenance information is surfaced per VM as well

Proactive-Redeploy to start maintenance

VM Experience – CRP

Initiate Maintenance – CRP

Planned Maintenance & Scheduled Events

Scheduled Events - API URI Description curl -H Metadata: true http: //169. 254/metadata/scheduledevents? api-version=2017 -03 -01 Retrieve all events for current tenant Response : { "Document. Incarnation": {Incarnation. ID}, "Events": [ { "Event. Id": {event. ID}, "Event. Type": "Reboot" | "Redeploy" | "Freeze", "Resource. Type": "Virtual. Machine", "Resources": [{resource. Name}], "Event. Status": "Scheduled" | "Started", "Not. Before": {time. In. UTC}, } ] } https: //docs. microsoft. com/en-us/azure/virtual-machines-scheduled-events Event Type • Pause – CPU is suspended, resources maintained ( e. g. active connections) • Reboot – VM Reboot, Memory wiped • Freeze – VM moved to another host. Ephemeral drive loss Event Status • Started – Events already approved and started • Scheduled – Events to be approved next

https: //myignite. microsoft. com/evaluations https: //aka. ms/ignite. mobileapp