Bug 2094260
Summary: | RHEL 9 kernel panic on Packer launched qemu-kvm | ||||||
---|---|---|---|---|---|---|---|
Product: | Red Hat Enterprise Linux 9 | Reporter: | Marko Myllynen <myllynen> | ||||
Component: | kernel | Assignee: | Virtualization Maintenance <virt-maint> | ||||
kernel sub component: | KVM | QA Contact: | liunana <nanliu> | ||||
Status: | CLOSED NOTABUG | Docs Contact: | |||||
Severity: | unspecified | ||||||
Priority: | unspecified | CC: | berrange, chayang, coli, darcari, dgilbert, imammedo, jinzhao, juzhang, nilal, prarit, virt-maint, vkuznets | ||||
Version: | 9.0 | ||||||
Target Milestone: | rc | ||||||
Target Release: | --- | ||||||
Hardware: | Unspecified | ||||||
OS: | Unspecified | ||||||
Whiteboard: | |||||||
Fixed In Version: | Doc Type: | If docs needed, set a value | |||||
Doc Text: | Story Points: | --- | |||||
Clone Of: | Environment: | ||||||
Last Closed: | 2022-06-27 09:49:04 UTC | Type: | Bug | ||||
Regression: | --- | Mount Type: | --- | ||||
Documentation: | --- | CRM: | |||||
Verified Versions: | Category: | --- | |||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||
Cloudforms Team: | --- | Target Upstream Version: | |||||
Embargoed: | |||||||
Attachments: |
|
Description
Marko Myllynen
2022-06-07 09:57:22 UTC
This is the corresponding issue at Packer GitHub: https://github.com/hashicorp/packer/issues/11839 Summarizing the above private discussion: The reported kernel panic is a result of using the default CPU model with qemu which is 'qemu64' which is deprecated. From Daniel's private comment above: Direct invocation of qemu-kvm is not supported in RHEL. We expect the user to launch the VM via libvirt, usually recommended with an mgmt tool which should only enable/use the supported option. For more details please refer to: https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/9/html/configuring_and_managing_virtualization/introducing-virtualization-in-rhel_configuring-and-managing-virtualization [quote] While QEMU is an essential component of the architecture, it is not intended to be used directly on RHEL 9 systems, due to security concerns. Therefore, using qemu-* commands is not supported by Red Hat, and it is highly recommended to interact with QEMU using libvirt. [/quote] We do have Bug#2060839 to deprecate CPU models like "kvm64" / "qemu64" on 9.1 from RHEL. Hence, I think we can close the bug as "NOTABUG" or as DUP of Bug#2060839. Any thoughts/concerns? Also, adding Igor to see if he has any additional thoughts. Thanks for looking into this. The above statement sounds good however in case if it's not possible to use libvirt (in all situations or in the short-term), would we have a suggestion how to invoke qemu-kvm to avoid kernel panic? I suspect using libvirt instead of qemu-kvm might require longer discussion and larger effort with Packer so I wonder do we have a temporary solution to allow Packer to build RHEL 9 images while they consider using libvirt? In any case, once we close this BZ I'll pass the info to the Packer/Qemu upstream ticket mentioned above. Thanks. (In reply to Marko Myllynen from comment #9) > Thanks for looking into this. > > The above statement sounds good however in case if it's not possible to use > libvirt (in all situations or in the short-term), would we have a suggestion > how to invoke qemu-kvm to avoid kernel panic? I suspect using libvirt > instead of qemu-kvm might require longer discussion and larger effort with > Packer so I wonder do we have a temporary solution to allow Packer to build > RHEL 9 images while they consider using libvirt? > > In any case, once we close this BZ I'll pass the info to the Packer/Qemu > upstream ticket mentioned above. > > Thanks. As you said in the initial comment, -cpu host works; that should generally be fine unless they plan live migration in a hetoregenous environment. (In reply to Dr. David Alan Gilbert from comment #10) > > As you said in the initial comment, -cpu host works; that should generally > be fine unless they plan live migration in a hetoregenous environment. Thanks for confirming. Packer definitely does not migrate anything, it's just an image building tool. Thanks, Daniel. Marko, are we good in closing this BZ now? If so, can you please in closing it? Thanks Thanks all, based on the above comments and the earlier BZ#2060839 I'm closing this as NOTABUG. I'll also provide a summary to the Packer upstream issue about this. |