euca-bundle-image fails to use symbolic links
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
euca2ools |
Fix Released
|
Undecided
|
Unassigned | ||
cloud-utils (Ubuntu) |
Fix Released
|
High
|
Scott Moser | ||
Lucid |
Fix Released
|
High
|
Scott Moser | ||
euca2ools (Ubuntu) |
Fix Released
|
High
|
Dustin Kirkland | ||
Lucid |
Fix Released
|
High
|
Dustin Kirkland |
Bug Description
While trying to run a lucid image on Lucid UEC the instance isn't able to be started.
According to the libvirt log file:
ubuntu@mabolo:~$ sudo cat /var/log/
LC_ALL=C PATH=/usr/
qemu: could not load kernel '/var/lib/
The following messages can be found in the kernel dmesg:
[324292.651948] type=1505 audit(126626372
[324292.661181] device vnet0 entered promiscuous mode
[324292.663037] br0: port 2(vnet0) entering learning state
[324292.664674] type=1503 audit(126626372
[324292.664695] type=1503 audit(126626372
[324292.961606] br0: port 2(vnet0) entering disabled state
[324293.039824] device vnet0 left promiscuous mode
[324293.039828] br0: port 2(vnet0) entering disabled state
[324322.741666] type=1505 audit(126626375
eucalyptus: 1.6.2~bzr1189-
libvirt: 0.7.5-5ubuntu7
qemu-kvm: 0.12.2-0ubuntu6
Changed in ubuntu: | |
status: | New → Incomplete |
summary: |
- Lucid UEC image 20100215 unable to start on UEC + euca-bundle-image fails to use symbolic links |
Changed in euca2ools (Ubuntu Lucid): | |
status: | New → Confirmed |
importance: | Undecided → Low |
Changed in euca2ools (Ubuntu Lucid): | |
assignee: | nobody → Dustin Kirkland (kirkland) |
importance: | Medium → High |
Changed in euca2ools: | |
status: | Fix Committed → Fix Released |
I suspect something is wrong before qemu-kvm is involved.
I just verified that I can do the following on 'mabolo':
=== go.sh ===
tmpd=$PWD/t
mkdir ${tmpd}
qemu-img create -f qcow2 "${tmpd}/disk" 1G
LC_ALL=C PATH=/usr/ local/sbin: /usr/local/ bin:/usr/ bin:/usr/ sbin:/sbin: /bin /usr/bin/kvm -M pc-0.12 -enable-kvm -m 512 -smp 2 -name i-4EC209E0 -uuid a4315a8f- 5c7c-ca4c- f158-4139e1663f d9 -nographic -chardev socket, id=monitor, path=${ tmpd}/i- 4EC209E0. monitor, server, nowait -monitor chardev:monitor -boot c -kernel ${1} -append "root=/dev/sda1 console=ttyS0" -drive file=${ tmpd}/disk, if=scsi, index=0, boot=on -net nic,macaddr= d0:0d:4e: c2:09:e0, vlan=0, model=e1000, name=e1000. 0 -net tap,fd= 65,vlan= 0,name= tap.0 -parallel none -usb -serial stdio
=== end go.sh ===
The differences in the command above are really only in where the monitor and are directed. With the above, kvm will successfully load the kernel and panic with "cant find root" (because there is no root in the newly created disk image).
I tested above with kernel lucid-server- uec-amd64- vmlinuz- virtual. 20100216