Score:0

What disk considerations are taken for VM migration during Host Maintenance in GCE?

in flag

I'm running a VM in the us-central1-a zone with Availability policies configured as follows:

Preemptibility - Off (Recommended) On host maintenance - Migrate VM instance (Recommended) Automatic restart - On (Recommended)

Today around 12 pm UTC-06:00 the VM was migrated. I'm using Dynatrace to monitor it and and at that time the machine raised a problem because it had went down.

In the GCP console the VM is shown as Running, I can ssh into it and my services are running just fine, however I can see that the Dynatrace agent or any of its files was not present in the VM hence the reason it was shown as down.

As the VM is from a dev environment it doesn't have snapshots configured so I wonder what is the process or what is considered when deciding what to include in the migration of the VM that doesn't have a snapshot, is it something random or is GCP taking snapshots on their own for this exact purpose?

Score:1
cn flag

What disk considerations are taken for VM migration during Host Maintenance in GCE?

There are no persistent disk considerations that you manage. Storage is not located on the hypervisor host. Migrations do not involve storage (snapshots, moving, restoring, etc.).

When Google performs a maintenance event, there are two major types: Host Maintenance and Lightweight.

Host maintenance performs kernel upgrades, hardware repair, etc on the machine running the hypervisor. Google uses live migration to move your VM to another hypervisor.

Lightweight performs hypervisor level upgrades, networking stack updates, etc. This does not require live migration.

Note: Live Migration does not take down your VM. This process should be transparent to your VM applications. If you are experiencing issues, open an issue tracker:

Report issues and request features with issue trackers

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.