[libvirt] qemu-kvm>=0.14 is unable to boot
Hello, I'm wondering what might be behind qemu-kvm>=0.14 is unable to boot except from PXE and perhaps virtio HDD. I mean, is somebody running qemu-kvm>=0.14 and libvirt-0.9.3/0.9.4 around here, thus is my setup broken? Thanks, Zdenek -- Zdenek Styblik email: stybla@turnovfree.net jabber: stybla@jabber.turnovfree.net
On 08/03/2011 01:16 AM, Zdenek Styblik wrote:
Hello,
I'm wondering what might be behind qemu-kvm>=0.14 is unable to boot except from PXE and perhaps virtio HDD.
I mean, is somebody running qemu-kvm>=0.14 and libvirt-0.9.3/0.9.4 around here, thus is my setup broken?
I had the problem that qemu-kvm-0.14 and libvirt-0.9.3 was unable to boot a *virtio* disk (which is exactly the opposite of what you type, but just in case that was a typo...). This was because qemu 0.14 was advertising (in the help) that bootindex was available, and libvirt saw that and told qemu to use it, but the version of seabios on my system (Fedora 14, running seabios-0.6.0) was too old to support it, so the boot failed. The solution was to add "<bootmenu enable='yes'/>" to the guest's config, which forces it to not use bootindex, and thus the guest will boot.
On 08/04/11 03:49, Laine Stump wrote:
On 08/03/2011 01:16 AM, Zdenek Styblik wrote:
Hello,
I'm wondering what might be behind qemu-kvm>=0.14 is unable to boot except from PXE and perhaps virtio HDD.
I mean, is somebody running qemu-kvm>=0.14 and libvirt-0.9.3/0.9.4 around here, thus is my setup broken?
I had the problem that qemu-kvm-0.14 and libvirt-0.9.3 was unable to boot a *virtio* disk (which is exactly the opposite of what you type, but just in case that was a typo...). This was because qemu 0.14 was advertising (in the help) that bootindex was available, and libvirt saw that and told qemu to use it, but the version of seabios on my system (Fedora 14, running seabios-0.6.0) was too old to support it, so the boot failed. The solution was to add "<bootmenu enable='yes'/>" to the guest's config, which forces it to not use bootindex, and thus the guest will boot.
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Laine, first of all, thanks. It did solved problem for me with libvirt-0.9.4 and qemu-kvm-0.15rc1. It is also solution for libvirt-0.9.3+ and qemu-kvm-0.14+. As for booting, I can honestly say whether virtio did boot or did not. And I have no way how to verify it right now(I have like one VM working so far in my new dev setup and still fiddling with kickstart). PXE did and didn't work, CD-ROM didn't, virtio ... I don't know. And perhaps it doesn't matter, because booting is broken. Is this "fix"/workaround noted somewhere? If not, it should be. And if told where/how, I volunteer to note it. As for seabios, I don't have one, or do I? I couldn't find it in packages/files, that's I'm sure of. Perhaps missing seabios is the problem? Zdenek -- Zdenek Styblik email: stybla@turnovfree.net jabber: stybla@jabber.turnovfree.net
On 08/04/11 20:16, Zdenek Styblik wrote:
On 08/04/11 03:49, Laine Stump wrote:
On 08/03/2011 01:16 AM, Zdenek Styblik wrote:
Hello,
I'm wondering what might be behind qemu-kvm>=0.14 is unable to boot except from PXE and perhaps virtio HDD.
I mean, is somebody running qemu-kvm>=0.14 and libvirt-0.9.3/0.9.4 around here, thus is my setup broken?
I had the problem that qemu-kvm-0.14 and libvirt-0.9.3 was unable to boot a *virtio* disk (which is exactly the opposite of what you type, but just in case that was a typo...). This was because qemu 0.14 was advertising (in the help) that bootindex was available, and libvirt saw that and told qemu to use it, but the version of seabios on my system (Fedora 14, running seabios-0.6.0) was too old to support it, so the boot failed. The solution was to add "<bootmenu enable='yes'/>" to the guest's config, which forces it to not use bootindex, and thus the guest will boot.
-- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
Laine,
first of all, thanks. It did solved problem for me with libvirt-0.9.4 and qemu-kvm-0.15rc1. It is also solution for libvirt-0.9.3+ and qemu-kvm-0.14+.
As for booting, I can honestly say whether virtio did boot or did not. And I have no way how to verify it right now(I have like one VM working so far in my new dev setup and still fiddling with kickstart). PXE did and didn't work, CD-ROM didn't, virtio ... I don't know. And perhaps it doesn't matter, because booting is broken. Is this "fix"/workaround noted somewhere? If not, it should be. And if told where/how, I volunteer to note it.
As for seabios, I don't have one, or do I? I couldn't find it in packages/files, that's I'm sure of. Perhaps missing seabios is the problem?
Zdenek
So I can confirm that booting off VirtIO disk. That is raw+LiLO and as far as finding boot sector and showing up LiLO. Then I'm getting kernel panic due to unable to open root-fs etc. Also booting off PXE works. However SCSI disk doesn't work without "hack" mentioned above. All with qemu-kvm-0.15rc1 and libvirt-0.9.4 without any "hacks". I'm not going to play with lower versions, but I'm confident it's going to be the same +/-. Z. -- Zdenek Styblik email: stybla@turnovfree.net jabber: stybla@jabber.turnovfree.net
participants (2)
-
Laine Stump -
Zdenek Styblik