openstack team mailing list archive

[Samuel Winchenbach <swinchen@xxxxxxxxx>]

I removed dynamic_ownership=0, restarted libvirtd-bin, qemu-kvm and all the
nova-* services.

I am still getting the same error: libvirtError: internal error Process
exited while reading console log output: chardev: opening backend "file"
failed: Permission denied


On Fri, Jun 21, 2013 at 2:19 PM, Samuel Winchenbach <swinchen@xxxxxxxxx>wrote:

> Also, I am curious  "Ensure that, NOVA-INST-DIR (set with state_path in
> nova.conf) is same on all hosts."
>
> should I set state_path=/os-grizzly/nova (My NFS mount) or leave it as
> /var/lib/nova on each node and set
> instances_path=/os-grizzly/nova/instances?
>
> It seems like each node should have it's own,unique state_path but I may
> be wrong.
>
> Sam
>
>
> On Fri, Jun 21, 2013 at 2:15 PM, Samuel Winchenbach <swinchen@xxxxxxxxx>wrote:
>
>> Ahhh a "bug" in the documentation?
>> http://docs.openstack.org/trunk/openstack-compute/admin/content/configuring-migrations.html
>>
>>
>> On Fri, Jun 21, 2013 at 2:14 PM, Daniel P. Berrange <berrange@xxxxxxxxxx>wrote:
>>
>>> On Fri, Jun 21, 2013 at 01:57:44PM -0400, Samuel Winchenbach wrote:
>>> > Here is some more information.  I am using Ubuntu 12.04 LTS, I have
>>> turned
>>> > off apparmor.
>>> >
>>> > Here is my mount: dedup:/big_pool/os-grizzly on /os-grizzly type nfs
>>> > (rw,noatime,nolock,tcp,bg,intr,hard,addr=10.54.90.10)
>>> >
>>> > root@test1:/# find /os-grizzly -type d | xargs ls -l -d
>>> > drwxr-xr-x 4 root   root   4 Jun 21 13:39 /os-grizzly
>>> > drwxr-xr-x 3 glance glance 3 Jun  3 10:25 /os-grizzly/glance
>>> > drwxr-x--- 2 glance glance 4 Jun 19 15:41 /os-grizzly/glance/images
>>> > drwxr-xr-x 3 nova   nova   3 Jun 21 13:39 /os-grizzly/nova
>>> > drwxr-xr-x 4 nova   nova   5 Jun 21 13:13 /os-grizzly/nova/instances
>>> > drwxr-xr-x 2 nova   nova   4 Jun 19 15:42
>>> /os-grizzly/nova/instances/_base
>>> > drwxr-xr-x 2 nova   nova   6 Jun 19 15:42
>>> /os-grizzly/nova/instances/locks
>>> >
>>> > root@test1:/# grep -RE "^[^#]" /etc/libvirt/*.conf
>>> > /etc/libvirt/libvirtd.conf:listen_tls = 0
>>> > /etc/libvirt/libvirtd.conf:listen_tcp = 1
>>> > /etc/libvirt/libvirtd.conf:unix_sock_group = "libvirtd"
>>> > /etc/libvirt/libvirtd.conf:unix_sock_rw_perms = "0770"
>>> > /etc/libvirt/libvirtd.conf:auth_unix_ro = "none"
>>> > /etc/libvirt/libvirtd.conf:auth_unix_rw = "none"
>>> > /etc/libvirt/libvirtd.conf:auth_tcp = "none"
>>> > /etc/libvirt/qemu.conf:dynamic_ownership = 0
>>> >
>>> >
>>> > I am stumped.  It was working fine before I made the changed
>>> > (dynamic_ownership, listen_tls, listen_tcp, auth_tcp) and started up
>>> the
>>> > other compute nodes.  :/
>>>
>>> Do not set 'dynamic_ownership' to 0. If you do this, you are required
>>> to have a mgmt app which knows how to set ownership on all resources
>>> used by QEMU. That option was added as a special hack for oVirt which
>>> can do that, but OpenStack does not support this.
>>>
>>> Daniel
>>> --
>>> |: http://berrange.com      -o-
>>> http://www.flickr.com/photos/dberrange/ :|
>>> |: http://libvirt.org              -o-
>>> http://virt-manager.org :|
>>> |: http://autobuild.org       -o-
>>> http://search.cpan.org/~danberr/ :|
>>> |: http://entangle-photo.org       -o-
>>> http://live.gnome.org/gtk-vnc :|
>>>
>>
>>
>