Re: [Openstack] Instance fails to spawn when instance_path is nfs mounted

Yong Sheng Gong Wed, 04 Apr 2012 04:21:50 -0700

Please have a look at my experience to see if it can help.

Thanks

-----openstack-bounces+gongysh=cn.ibm....@lists.launchpad.net wrote: -----

To: Razique Mahroua <razique.mahr...@gmail.com>
From: Yong Sheng Gong/China/IBM@IBMCN
Sent by: openstack-bounces+gongysh=cn.ibm....@lists.launchpad.net
Date: 03/29/2012 07:39PM
Cc: "openstack@lists.launchpad.net $openstack@lists.launchpad.net$" <openstack@lists.launchpad.net>
Subject: Re: [Openstack] KVM cannot start vm when the disk is on NFS

Thanks for the reply.

I have figured it out.

I am running nova control node with root account on the side of nfs4 server side (redhat6.1), while on the side of compute node is ubuntu.

So on ubuntu side, I have executed:
$sudo bash,
#nova-compute

In fact what I need to do is to let kvm binary run under root user id.
I found the solution from http://libvirt.org/drvqemu.html.

Here are what I did on ubuntu side:
#chown root:root /var/run/libvirt/qemu /var/lib/libvirt/qemu /var/cache/libvirt/qemu

edit /etc/libvirt/qemu.conf
# The user ID for QEMU processes run by the system instance.
user = "root"

# The group ID for QEMU processes run by the system instance.
group = "root"

after restarting libvirt-bin, nova-compute can provision vm from the images on NFS share dir now.

-----Razique Mahroua <razique.mahr...@gmail.com> wrote: -----

To: Yong Sheng Gong/China/IBM@IBMCN
From: Razique Mahroua <razique.mahr...@gmail.com>
Date: 03/29/2012 06:00PM
Cc: "openstack@lists.launchpad.net (openstack@lists.launchpad.net)" <openstack@lists.launchpad.net>
Subject: Re: [Openstack] KVM cannot start vm when the disk is on NFS

Hello Yong,
as you may understand, the chmod 777 is not the best thing to do.
it looks like the files belong to the user root, while it's the user nova which runs the cloud.

The first thing to do is to give the rights to the user nova of the directory (755) to nova:nova.
Thanks!

Nuage & Co - Razique Mahroua
razique.mahr...@gmail.com

Le 29 mars 2012 à 09:50, Yong Sheng Gong a écrit :

I have to run chmod 777 /var/lib/nova/instances/instance-00000007 to start the vm on kvm. Appreciate any help from stackers.

HostA ( 192.168.1.90): redhat 6.1 nfs4
/etc/exports:
/var/lib/nova/instances 192.168.1.0/24(rw,sync,fsid=0,no_root_squash)

/etc/idmapd:
[General]
Domain = openstack

#service nfs restart
#service rpcidmapd restart

HostB ( 192.168.1.94): ubuntu 11.10
#sudo bash

/etc/idmapd:
[General]

Verbosity = 0
Pipefs-Directory = /var/lib/nfs/rpc_pipefs
Domain = openstack
/etc/fstab:
192.168.1.90:/ /var/lib/nova/instances nfs4 defaults 0 0

#mount -a -v
root@openubuntu:/var/lib/nova/instances/instance-00000007# ll
total 14168
drwxr-xr-x 2 root   root       4096 2012-03-29 14:23 ./
drwxr-xr-x 9 nobody nogroup    4096 2012-03-29 14:21 ../
-rw-rw---- 1 root   root          0 2012-03-29 14:17 console.log
-rw-r--r-- 1 root   root    6291968 2012-03-29 14:18 disk
-rw-r--r-- 1 root   root    4404752 2012-03-29 14:18 kernel
-rw-r--r-- 1 root   root       1619 2012-03-29 14:17 libvirt.xml
-rw-r--r-- 1 root   root    5882349 2012-03-29 14:18 ramdisk
root@openubuntu:/var/lib/nova/instances/instance-00000007# virsh start instance-00000007
error: Failed to start domain instance-00000007
error: internal error Process exited while reading console log output: chardev: opening backend "file" failed

root@openubuntu:/var/lib/nova/instances/instance-00000007# chmod 777 console.log
root@openubuntu:/var/lib/nova/instances/instance-00000007# virsh start instance-00000007
error: Failed to start domain instance-00000007
error: internal error Process exited while reading console log output: char device redirected to /dev/pts/3
kvm: -drive file=/var/lib/nova/instances/instance-00000007/disk,if=none,id=drive-virtio-disk0,format=qcow2,cache=none: could not open disk image /var/lib/nova/instances/instance-00000007/disk: Permission denied

root@openubuntu:/var/lib/nova/instances/instance-00000007# chmod 777 *
root@openubuntu:/var/lib/nova/instances/instance-00000007# virsh start instance-00000007
Domain instance-00000007 started

Thanks
Yong Sheng Gong

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to     : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help : https://help.launchpad.net/ListHelp

-----openstack-bounces+gongysh=cn.ibm....@lists.launchpad.net wrote: -----

To: Nathanael Burton <nathanael.i.bur...@gmail.com>, Diego Parrilla Santamaría <diego.parrilla.santama...@gmail.com>
From: "Vaze, Mandar" <mandar.v...@nttdata.com>
Sent by: openstack-bounces+gongysh=cn.ibm....@lists.launchpad.net
Date: 04/04/2012 06:28PM
Cc: "openstack@lists.launchpad.net" <openstack@lists.launchpad.net>
Subject: Re: [Openstack] Instance fails to spawn when instance_path is nfs mounted

> I had a problem like this when the umask was locked down. Setting the umask to 022 in the init script for nova-compute solved my problem.

I'm running from devstack setup - so I don't have an init-script. All the processes run in a "screen" session.
Here is what I did :

1. Stopped nova-compute using Ctrl-C
2. Checked existing umask using "umask command" it was set to 0002
3. Updated umask using "umask 022" - checked that it was indeed updated - this time "umask" command returned 0022
4. Started nova-compute as follows :
mandar@ubuntu-dev-mandar:/opt/stack/nova$ cd /opt/stack/nova && sg libvirtd /opt/stack/nova/bin/nova-compute
5. Spawned an instance from Dashboard

Still getting same error.

-Mandar

On Tue, Apr 3, 2012 at 1:56 PM, Diego Parrilla Santamaría <diego.parrilla.santama...@gmail.com> wrote:
> We use nfs backed instances a lot, and this problem normally has to do
> with wrong permission management in your filer and/or client.
>
> Check if not only root can write on the nfs share (specially libvirt user).
>
> Diego
> --
> Diego Parrilla
> CEO
> www.stackops.com | diego.parri...@stackops.com | +34 649 94 43 29
> | skype:diegoparrilla
>
>
>
>
> On Tue, Apr 3, 2012 at 6:43 PM, Mandar Vaze / मंदार वझे
> <mandarv...@gmail.com> wrote:
>>
>> I saw an old question posted here :
>> https://answers.launchpad.net/nova/+question/164689
>>
>> But I am not trying live migration.
>>
>> I have nfs mounted instances_path - so when I try to spawn an
>> instance I run into the above errors. Especially following :
>>
>> File "/usr/lib/python2.7/dist-packages/libvirt.py", line 372, in
>> createWithFlags
>> 40842 2012-04-03 05:42:27 TRACE nova.rpc.amqp if ret == -1: raise
>> libvirtError ('virDomainCreateWithFlags() failed', dom=self)
>> 40843 2012-04-03 05:42:27 TRACE nova.rpc.amqp libvirtError: internal
>> error Process exited while reading console log output: chardev:
>> opening backend "file" failed
>>
>> But as you can see below, several files are created in this folder,
>> so I am not sure if mine if permissions issue (Else none of the files
>> would get created) The problem is reported when libvirt tries to
>> write to console.log (File itself is created with correct permissions
>> - just that this is zero byte file)
>>
>> mandar@ubuntu-dev-mandar:/nfs_shared_instances_path/instance-00000005
>> $ ll total 10944 drwxrwxr-x 2 mandar libvirtd 4096 2012-04-03
>> 05:42 ./ drwxrwxrwx 4 root root 4096 2012-04-03 05:42 ../
>> -rw-rw---- 1 mandar libvirtd 0 2012-04-03 05:42 console.log
>> -rw-r--r-- 1 mandar libvirtd 6291968 2012-04-03 05:42 disk
>> -rw-rw-r-- 1 mandar libvirtd 4731440 2012-04-03 05:42 kernel
>> -rw-rw-r-- 1 mandar libvirtd 1067 2012-04-03 05:42 libvirt.xml
>> -rw-rw-r-- 1 mandar libvirtd 2254249 2012-04-03 05:42 ramdisk
>>
>> I'm suspecting :
>> https://bugs.launchpad.net/ubuntu/maverick/+source/libvirt/+bug/63269
>> 6 But I the above doesn't show itself in non-NFS setup
>>
>> Please suggest !!!
>>
>> -Mandar
>>
>>
>> _______________________________________________
>> Mailing list: https://launchpad.net/~openstack Post to :
>> openstack@lists.launchpad.net Unsubscribe :
>> https://launchpad.net/~openstack More help :
>> https://help.launchpad.net/ListHelp
>>
>
>
> _______________________________________________
> Mailing list: https://launchpad.net/~openstack Post to :
> openstack@lists.launchpad.net Unsubscribe :
> https://launchpad.net/~openstack More help :
> https://help.launchpad.net/ListHelp
>

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help : https://help.launchpad.net/ListHelp

______________________________________________________________________
Disclaimer:This email and any attachments are sent in strictest confidence for the sole use of the addressee and may contain legally privileged, confidential, and proprietary data. If you are not the intended recipient, please advise the sender by replying promptly to this email and then delete and destroy this email and any attachments without any further use, copying or forwarding
_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help : https://help.launchpad.net/ListHelp

_______________________________________________
Mailing list: https://launchpad.net/~openstack
Post to     : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp

Re: [Openstack] Instance fails to spawn when instance_path is nfs mounted

Reply via email to