feat(xo-server): implement rolling pool reboot

[fbeauchamp]

Description

Short explanation of this PR (feel free to re-use commit message)

Checklist

• Commit
  • Title follows commit conventions
  • Reference the relevant issue (Fixes #007, See xoa-support#42, See https://...)
  • If bug fix, add Introduced by
• Changelog
  • If visible by XOA users, add changelog entry
  • Update "Packages to release" in CHANGELOG.unreleased.md
• PR
  • If UI changes, add screenshots
  • If not finished or not tested, open as Draft

[MathieuRA]

Since we are very close to the release, what do you think if for the rolling Pool Reboot we just copy/paste the necessary code, and think about how to mutualise the code right after the release?

[fbeauchamp]

Since we are very close to the release, what do you think if for the rolling Pool Reboot we just copy/paste the necessary code, and think about how to mutualise the code right after the release?

if possible, I would like to try the refactoring, the rolling pool update method is already complex enough . If we're not confident on the code quality , we can still fallback to an additionnal flag on wednesday

[julien-f]

I would love the logic to be mutualized with host_smartReboot as well.

[fbeauchamp]

I would love the logic to be mutualized with host_smartReboot as well.

how do you see things ? since the code logic seems quite different , I think I need more time to think this out. Each Vm will be handled exactly twice in smartreboot ( suspend, then resume), but an unlucky VM can be migrated a lot of times in case of rolling pull reboot . Also the error cases are different (allowed operation Vs local storage)

I think both feature should be blended : in case of rolling pool update/reboot , we could fall back to suspend for example

[MathieuRA]

Not tested.

[MathieuRA]

Don't forget to rebase on master to get the "Rolling pool update" bug fix.

// host.$resident_VMs is outdated and returns resident VMs before the host.evacuate.
// this.getField is used in order not to get cached data.
const residentVmRefs = await this.getField('host', host.$ref, 'resident_VMs')

[fbeauchamp]

test done : rolling reboot et rolling pull update works