|Summary:||[CNV][Chaos] Ensure MHC is enabled before starting VMs that are required to be highly available|
|Product:||Container Native Virtualization (CNV)||Reporter:||Piotr Kliczewski <pkliczew>|
|Component:||Installation||Assignee:||Erkan Erol <eerol>|
|Status:||NEW ---||QA Contact:||Inbar Rose <irose>|
|Version:||future||CC:||abeekhof, cnv-qe-bugs, danken, msluiter, racedoro, shardy, stirabos|
|Fixed In Version:||Doc Type:||If docs needed, set a value|
|Doc Text:||Story Points:||---|
|oVirt Team:||---||RHEL 7.3 requirements from Atomic Host:|
|Cloudforms Team:||---||Target Upstream Version:|
|Bug Depends On:|
Description Piotr Kliczewski 2020-10-20 10:15:40 UTC
In order to make sure chaos scenarios won't affect user workload we need to enable machine health check by default on freshly installed clusters.
Comment 1 Steven Hardy 2020-10-20 11:54:43 UTC
This was previously discussed (and rejected) via https://github.com/openshift/enhancements/pull/330 It sounds like we need to find a way to either enable users to easily opt-in when they want this behavior, or for CNV to automate that as part of the CNV setup In either case the baremetal-operator is the wrong component for this, so I'll defer triaging until we can decide if this bug should be closed or moved to a different component.
Comment 2 Rob Young 2020-10-20 15:26:21 UTC
@shardy we need to revisit the larger platform problem that our current CNV negative flow testing has exposed, specifically around BM deployments. I propose leaving this bug open until we have a proper hearing on the findings and our options for a platform level solution. We don't want to create a one-off solution specific to CNV; the platform behavior and UX should be consistent regardless of feature that exposes the issue.
Comment 5 Steven Hardy 2020-10-22 14:28:24 UTC
(In reply to Rob Young from comment #2) > @shardy we need to revisit the larger platform problem that our current CNV > negative flow testing has exposed, specifically around BM deployments. I > propose leaving this bug open until we have a proper hearing on the findings > and our options for a platform level solution. We don't want to create a > one-off solution specific to CNV; the platform behavior and UX should be > consistent regardless of feature that exposes the issue. Ack, we can leave it open, but we need to find a different component I think - the baremetal-operator is unrelated to MHC and it seems the solution is likely to be tracked via CNV teams?