xenapi: ImageTooLarge for instances with 0 disk size

[Mate Lakat]

Create two flavors:

nova-manage flavor create --name=m1.nodisk --memory=512 --cpu=1 --root_gb=0 \
--flavor=7 --swap=0 --rxtx_factor=1.0 --is_public=true
nova-manage flavor create --name=m1.onegig --memory=512 --cpu=1 --root_gb=1 \
--flavor=8 --swap=0 --rxtx_factor=1.0 --is_public=true

Upload an image:

glance image-create --name growbug \
--copy-from=https://github.com/downloads/citrix-openstack/warehouse/cirros-0.3.0-x86_64-disk.vhd.tgz \
--container-format=ovf --disk-format=vhd --is-public=True

Wait until it is active, get its id:

IMAGEID=$(glance image-list | grep bugimage | sed -e 's/^| //g' -e 's/ |.*$//g')

Boot it up:

nova boot --image $IMAGEID --flavor 7 growinstance

The instance state will be error:

stack@DevStackOSDomU:~/devstack$ nova list | grep growinstance
| 200fd365-6d7d-45b0-9c35-89ef11d43662 | growinstance | ERROR | |

And you see the following in the n-cpu log:

Failed to spawn, rolling back
Traceback (most recent call last):
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 452, in spawn
    vdis = create_disks_step(undo_mgr, disk_image_type, image_meta)
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 143, in inner
    rv = f(*args, **kwargs)
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 335, in create_disks_step
    block_device_info=block_device_info)
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 306, in _create_disks
    block_device_info=block_device_info)
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 518, in get_vdis_for_instance
    context, session, instance, name_label, image, image_type)
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 974, in _create_image
    image_id, image_type)
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 919, in _create_cached_image
    image_id, image_type)
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 994, in _fetch_image
    vdis = _fetch_vhd_image(context, session, instance, image_id)
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 1120, in _fetch_vhd_image
    _check_vdi_size(context, session, instance, root_vdi_uuid)
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 1157, in _check_vdi_size
    raise exception.ImageTooLarge()
ImageTooLarge: Image is larger than instance type allows

[Mate Lakat (mate-lakat)]

...

Now, start an instance with a non-zero flavor:
stack@DevStackOSDomU:~/devstack$ nova flavor-list | grep m1.onegig
| 8 | m1.onegig | 512 | 1 | 0 | | 1 | 1.0 | True | {} |

Start it:

nova boot --image $IMAGEID --flavor 8 nonzeroinstance

The instance will be booted:

stack@DevStackOSDomU:~/devstack$ nova list | grep nonzeroinstance
| afda8249-8026-4c6f-a577-3866206606c1 | nonzeroinstance | ACTIVE | private=10.0.0.5 |

Now destroy the non-zero:

nova delete afda8249-8026-4c6f-a577-3866206606c1

And try to launch an instance with auto-disk size:

stack@DevStackOSDomU:~/devstack$ nova flavor-list | grep m1.nodisk
| 7 | m1.nodisk | 512 | 0 | 0 | | 1 | 1.0 | True | {} |

nova boot --image $IMAGEID --flavor 7 growinstance

That will fail in a different way:

Failed to spawn, rolling back
Traceback (most recent call last):
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 456, in spawn
    attach_disks_step(undo_mgr, vm_ref, vdis, disk_image_type)
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 143, in inner
    rv = f(*args, **kwargs)
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 388, in attach_disks_step
    injected_files)
  File "/opt/stack/nova/nova/virt/xenapi/vmops.py", line 568, in _attach_disks
    instance_type['root_gb'])
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 769, in auto_configure_disk
    _resize_part_and_fs(dev, start, old_sectors, new_sectors)
  File "/opt/stack/nova/nova/virt/xenapi/vm_utils.py", line 2032, in _resize_part_and_fs
    run_as_root=True)
  File "/opt/stack/nova/nova/utils.py", line 243, in execute
    cmd=' '.join(cmd))
ProcessExecutionError: Unexpected error while running command.
Command: sudo nova-rootwrap /etc/nova/rootwrap.conf resize2fs /dev/xvdb1 -16065s
Exit code: 1
Stdout: ''
Stderr: "resize2fs 1.42 (29-Nov-2011)\n/sbin/resize2fs: invalid option -- '1'\nUsage: /sbin/resize2fs [-d debug_flags] [-f] [-F] [-M] [-P] [-p] device [new_size]\n\n"

[Chuck Short (zulcss)]

Which version is this with?

[Mate Lakat (mate-lakat)]

It was on OS - trunk, XenServer 6.1

[John Garbutt (johngarbutt)]

There are many issues with disksize 0 flavour.

[John Garbutt (johngarbutt)]

this is related: https://bugs.launchpad.net/nova/+bug/1152268

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/29848
Committed: http://github.com/openstack/nova/commit/f692f727236b7e3e97b657c048194870ddf88a47
Submitter: Jenkins
Branch: master

commit f692f727236b7e3e97b657c048194870ddf88a47
Author: Rick Harris <email address hidden>
Date: Mon May 20 23:51:09 2013 +0000

    xenapi: Disable VDI size check when root_gb=0

    Nova uses root_gb=0 to indicate that we should use the image as-is without
    trying to resize the root disk. Currently the xenapi virt-driver doesn't
    properly implement this behavior.

    This patch gets us a step closer by disabling the VDI size check when
    root_gb=0. We still need to ensure that the disk-resize code handle root_gb=0
    as well.

    References bug 1155113

    Change-Id: If91963e079e197a1315c48b8643f3419b42b0b2c