Not checking image format produces lots of useless messages

[Stanislaw Pitucha]

The code for resizing partitionless images goes with the "tell, don't ask" idea and attempts to run extfs / mount utilities on an image even though they may fail. This produces lots of useless messages during the instance preparation, like these:

2013-11-21 06:45:07 20902 DEBUG nova.virt.disk.mount.api [req-939d5d50-25ea-4f7a-8882-a880b9671e47 10873781609182 10816527907643] Failed to mount filesystem: Unexpected error while running command.
Command: sudo nova-rootwrap /etc/nova/rootwrap.conf mount /dev/nbd8 /tmp/openstack-vfs-localfsSz6ylg
Exit code: 32
Stdout: ''
Stderr: 'mount: you must specify the filesystem type\n' mnt_dev /usr/lib/python2.7/dist-packages/nova/virt/disk/mount/api.py:198
2013-11-21 06:45:07 20902 DEBUG nova.virt.disk.mount.api [req-939d5d50-25ea-4f7a-8882-a880b9671e47 10873781609182 10816527907643] Fail to mount, tearing back down do_mount /usr/lib/python2.7/dist-packages/nova/virt/disk/mount/api.py:219
2013-11-21 06:45:07 20902 DEBUG nova.virt.disk.mount.api [req-939d5d50-25ea-4f7a-8882-a880b9671e47 10873781609182 10816527907643] Unmap dev /dev/nbd8 unmap_dev /usr/lib/python2.7/dist-packages/nova/virt/disk/mount/api.py:184
2013-11-21 06:45:07 20902 DEBUG nova.virt.disk.mount.nbd [req-939d5d50-25ea-4f7a-8882-a880b9671e47 10873781609182 10816527907643] Release nbd device /dev/nbd8 unget_dev /usr/lib/python2.7/dist-packages/nova/virt/disk/mount/nbd.py:128
2013-11-21 06:45:07 20902 DEBUG nova.openstack.common.processutils [req-939d5d50-25ea-4f7a-8882-a880b9671e47 10873781609182 10816527907643] Running cmd (subprocess): sudo nova-rootwrap /etc/nova/rootwrap.conf qemu-nbd -d /dev/nbd8 execute /usr/lib/python2.7/dist-packages/nova/openstack/common/processutils.py:147
2013-11-21 06:45:07 20902 DEBUG nova.virt.disk.vfs.localfs [req-939d5d50-25ea-4f7a-8882-a880b9671e47 10873781609182 10816527907643] Failed to mount image Failed to mount filesystem: Unexpected error while running command.
Command: sudo nova-rootwrap /etc/nova/rootwrap.conf mount /dev/nbd8 /tmp/openstack-vfs-localfsSz6ylg
Exit code: 32
Stdout: ''
Stderr: 'mount: you must specify the filesystem type\n') setup /usr/lib/python2.7/dist-packages/nova/virt/disk/vfs/localfs.py:83
2013-11-21 06:45:07 20902 DEBUG nova.virt.disk.api [req-939d5d50-25ea-4f7a-8882-a880b9671e47 10873781609182 10816527907643] Unable to mount image /var/lib/nova/instances/5159cff0-91f6-4521-a0be-d74ce4f81fad/disk with error Failed to mount filesystem: Unexpected error while running command.
Command: sudo nova-rootwrap /etc/nova/rootwrap.conf mount /dev/nbd8 /tmp/openstack-vfs-localfsSz6ylg
Exit code: 32
Stdout: ''
Stderr: 'mount: you must specify the filesystem type\n'. Cannot resize. is_image_partitionless /usr/lib/python2.7/dist-packages/nova/virt/disk/api.py:191

This could be fixed by doing a simple check on the the image itself to pick up the magic signature. This would allow to skip e2resize on non-extfs files and simple mounting on partitioned images without the error messages.

[[e-mail] (liuh101410)]

I had the same error, when I make snapshot of a running instance.

 Failed to mount filesystem: Unexpected error while running command.
Command: sudo nova-rootwrap /etc/nova/rootwrap.conf mount /dev/nbd14 /tmp/openstack-vfs-localfs4ZzQmW
Exit code: 32
Stdout: ''
Stderr: 'mount: you must specify the filesystem type\n' mnt_dev /usr/lib/python2.7/dist-packages/nova/virt/disk/mount/api.py:198

When I fix the error:
 error: unable to execute QEMU command 'drive-mirror': Could not open '/var/lib/nova/instances/snapshots/tmpNtVLgi/25c19f3475ae4fec8f0dd3bb36dda2e4.delta'

I added the following line to /etc/apparmor/abstractions/libvirt-qemu
/var/lib/nova/instances/snapshots/*/*.delta rw,

And then, the first error happened.
I think this problem may caused by apparmor. Because when I removed it , I can make snapshot of a running instance successfully. But when I used NFS as shared storage, apparmor must be used.

[[e-mail] (liuh101410)]

sorry, the path is /etc/apparmor.d/abstractions/libvirt-qemu

[jichenjc (jichenjc)]

I did see the function is_image_partitionless has check logic

if it's qcow2, it will try to mount the device through qemu-nbd which can't use e2label
if it's a raw, it will try to e2label to check whether it's a ext disk or not

[jichenjc (jichenjc)]

so I will not change the code logic, instead, will add a silent flag to ignore debug error info in
disk/api.py is_image_partitionless function for
fs.setup()
fs.teardown()
function

[jichenjc (jichenjc)]

did see a review
https://review.openstack.org/#/c/64383/
don't know whether it's related to this

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/69896

[jichenjc (jichenjc)]

according to review comments
https://review.openstack.org/#/c/69896/

we might keep the log info since it will have lot of useful info

[Carmelo Ragusa (carmelo-ragusa)]

has this been abandoned? I'm suddenly getting this error and this is the only place where there was some discussion. I thought it was fixed but I checked the review comments above and doens't seem fixed to me. Any idea?