XML

Word

Printable

Type: Epic
Resolution: Unresolved
Priority: Major
Fix Version/s: None
Affects Version/s: None
Component/s: CNV Install, Upgrade and Operators
Labels:
- 418
- CY25
- cnv-observability
- fabiand-shortlist
- grooming
- no-ui

Epic Name:
best-effort-eviction
Blocked:
False
Blocked Reason:

Hide

None

Show
None
Ready:
False
Acceptance Criteria:
Hide

MUST evaluate if PDBs are still the right way to block VMs, as PDBS ultimately block the upgrade

MUST provide a best effort startegy for eventually forcing upgrades to roll through

MUST address plain evictions

MUST address evictions as part of upgrade flows
Show
MUST evaluate if PDBs are still the right way to block VMs, as PDBS ultimately block the upgrade MUST provide a best effort startegy for eventually forcing upgrades to roll through MUST address plain evictions MUST address evictions as part of upgrade flows
Epic Status:
To Do
Feature Link:
CNV-8280 - VM balancing
Hierarchy Progress Bar:

73% To Do, 9% In Progress, 18% Done
Ready-Ready:

doc-ready, ux-ready
[QE] How to address?:
---
[QE] Why QE missed?:
---
Target Version:

CNV v4.18.0

SFDC Cases Links:
SFDC Cases Counter:
SFDC Cases Open:

Market:

Goal

virt-launcher pods are evicted during CNV and OCP updates. When live migration is used during eviction, it's possible to block the eviction indefinitely if live migration fails continuously.

The goal is to provide a cluster scoped tuning that treats live migration during updates (Both OCP and CNV) as best effort. Best effort means that live migration should be attempted multiple times, but eventually give up and fall back to a restart.

User Stories

As a CNV cluster admin, during a CNV update I want to ensure all VMI workloads are eventually updated with new virt-launchers without intervention in the event that live migration of a VMI continually fails.
As an OCP cluster admin, I don't want it to be possible for KubeVirt virtual machines to block OCP node updates indefinitely due to failed live migration.

Non-Requirements

Notes

Any additional details or decisions made/needed

Done Checklist

Who	What	Reference
DEV	Upstream roadmap issue (or individual upstream PRs)	<link to GitHub Issue>
DEV	Upstream documentation merged	<link to meaningful PR>
DEV	gap doc updated	<name sheet and cell>
DEV	Upgrade consideration	<link to upgrade-related test or design doc>
DEV	CEE/PX summary presentation	label epic with cee-training and add a <link to your support-facing preso>
QE	Test plans in Polarion	<link or reference to Polarion>
QE	Automated tests merged	<link or reference to automated tests>
DOC	Downstream documentation merged	<link to meaningful PR>

is related to

CNV-35883 Enable defining schedule/acks/tuning for workloadUpdateStrategy

relates to

CNV-35883 Enable defining schedule/acks/tuning for workloadUpdateStrategy

CNV-38291 For vms that can't be livemigrated due to lack of target node with matching resources eviction strategy LiveMigrateIfPossible does not prioritizes upgrades, and blocks upgrades

CNV-38193 No ability to evict a hotplugged vm this is blocking ocp upgrade

Closed

split from

CNV-28528 Enhance Descheduler to support VMs

Closed

Assignee:: Stuart Gott

Reporter:: David Vossel

QA Contact:: Debarati Basu-Nag

Votes:: 0 Vote for this issue

Watchers:: 9 Start watching this issue

Created:: 2023/11/13 10:03 PM

Updated:: 2024/10/23 9:49 AM

Details

Description

Goal

User Stories

Non-Requirements

Notes

Done Checklist

Attachments

Issue Links

Activity

People

Dates