[one-users] Help with datastore feature

Mon May 14 07:11:06 PDT 2012

Hi Robert,

Let's see if you can pinpoint where the problem is. A few ideas you may check:

- Has the 'oneadmin' user the same unix id across all the servers?

- You said in your email that "ownership change is disabled" however
in the documentation [1] it states that "To be able to use the images
copied by OpenNebula, change also the user and group under which the
libvirtd is run to “oneadmin”." Is it properly configured?

- You could prevent OpenNebula from removing the images after a fail,
so you can do a manual "virsh create":
1. [frontend] comment the ssh_exec_and_log $DST_HOST "rm -rf ..." line
in /var/lib/one/remotes/tm/<TM_MAD>/delete
2. [host] cd /var/lib/one/datastores/0/<VM_ID>; virsh create deployment.0
3. [host] ls -l `readlink -f /var/lib/one/datastores/0/<VM_ID>/disk.0`
4. Are you able to manipulate as oneadmin the
/var/lib/one/datastores/0/<VM_ID>/disk.0 image? try doing echo "foo" >
/path/to/image

[1] http://opennebula.org/documentation:rel3.4:kvmg

Regards,
Jaime

On Sat, May 12, 2012 at 1:44 AM, Robert Schweikert <rjschwei at suse.com> wrote:
>
> On 05/11/2012 04:17 PM, Ruben S. Montero wrote:
>>
>> Hi Robert,
>>
>>  From your output it seems that the problem is probably at the libvirt
>> configuration. In particular:
>>
>> Fri May 11 14:48:00 2012 [VMM][I]: WARNING: no socket to connect to
>
>
> Well, I have made no changes from previous working versions w.r.t. libvirtd configuration.
>
>
>>
>> Could you double check the configuration hints at
>>
>> http://www.opennebula.org/documentation:rel3.4:kvmg#kvm_configuration
>
>
> # cat /etc/polkit-1/localauthority/50-local.d/60-suseNebula-access.pkla [Remote libvirt SSH access]
>  Identity=unix-user:oneadmin
>  Action=org.libvirt.unix.manage
>  ResultAny=yes
>  ResultInactive=yes
>  ResultActive=yes
>
> Which matches the config given in the doc.
>
> - tcp listening is enabled
> # grep tcp /etc/libvirt/libvirtd.conf
> listen_tcp = 1
> #tcp_port = "16509"
> #auth_tcp = "sasl"
>
> - ownership change is disabled
> # grep owner /etc/libvirt/qemu.conf
> # Whether libvirt should dynamically change file ownership
> # Set to 0 to disable file ownership changes.
> dynamic_ownership = 0
>
> - On SUSE distribution libvirtd is started with -l command line option, thus there is no need for modification of /etc/default/libvirt-bin
> # grep startproc /etc/init.d/libvirtd
>        startproc $LIBVIRTD_BIN -d -l
>
> - The oneadmin user is part of the kvm, root, and cloud groups (on SUSE there is no libvirtd group, everything w.r.t. libvirtd is part of root).
> # grep oneadmin /etc/group
> root:x:0:oneadmin
> kvm:!:105:qemu,oneadmin
> cloud:!:1000:oneadmin
>
> Thus the libvirtd configuration appears to be OK. I switched the driver from qcow2 to shared, the /var/lib/one directory is nfs mounted on the cloud node
> # mount | grep one
> 192.168.1.202:/var/lib/one on /var/lib/one type nfs4 (rw,relatime,vers=4,rsize=262144,wsize=262144,namlen=255,hard,proto=tcp,port=0,timeo=600,retrans=2,sec=sys,clientaddr=192.168.1.203,minorversion=0,local_lock=none,addr=192.168.1.202)
>
> But I still appears to have a permission problem
>  cat /var/log/one/0.log
> Fri May 11 19:25:09 2012 [DiM][I]: New VM state is ACTIVE.
> Fri May 11 19:25:09 2012 [LCM][I]: New VM state is PROLOG.
> Fri May 11 19:25:09 2012 [VM][I]: Virtual Machine has no context
> Fri May 11 19:25:32 2012 [TM][I]: clone: Cloning ../../1/2b7b3d62a1684540409e2073f3b0b036 in 192.168.1.203:/var/lib/one//datastores/0/0/disk.0
> Fri May 11 19:25:32 2012 [TM][I]: ExitCode: 0
> Fri May 11 19:25:32 2012 [LCM][I]: New VM state is BOOT
> Fri May 11 19:25:32 2012 [VMM][I]: Generating deployment file: /var/lib/one/0/deployment.0
> Fri May 11 19:25:33 2012 [VMM][I]: ExitCode: 0
> Fri May 11 19:25:33 2012 [VMM][I]: Successfully execute network driver operation: pre.
> Fri May 11 19:25:33 2012 [VMM][I]: Command execution fail: cat << EOT | /var/tmp/one/vmm/kvm/deploy /var/lib/one//datastores/0/0/deployment.0 192.168.1.203 0 192.168.1.203
> Fri May 11 19:25:33 2012 [VMM][I]: WARNING: no socket to connect to
> Fri May 11 19:25:33 2012 [VMM][I]: error: Failed to create domain from /var/lib/one//datastores/0/0/deployment.0
> Fri May 11 19:25:33 2012 [VMM][I]: error: internal error process exited while connecting to monitor: kvm: -drive file=/var/lib/one//datastores/0/0/disk.0,if=none,id=drive-ide0-0-0,format=qcow2: could not open disk image /var/lib/one//datastores/0/0/disk.0: Permission denied
> Fri May 11 19:25:33 2012 [VMM][I]:
> Fri May 11 19:25:33 2012 [VMM][E]: Could not create domain from /var/lib/one//datastores/0/0/deployment.0
> Fri May 11 19:25:33 2012 [VMM][I]: ExitCode: 255
> Fri May 11 19:25:33 2012 [VMM][I]: Failed to execute virtualization driver operation: deploy.
> Fri May 11 19:25:33 2012 [VMM][E]: Error deploying virtual machine: Could not create domain from /var/lib/one//datastores/0/0/deployment.0
> Fri May 11 19:25:34 2012 [DiM][I]: New VM state is FAILED
>
> The disk.0 file exists:
> ls -l /var/lib/one/datastores/0/0
> total 560428
> -rw-r--r-- 1 oneadmin cloud       630 May 11 23:25 deployment.0
> -rw-r----- 1 oneadmin cloud 573308928 May 11 23:25 disk.0
>
> When I changed the mode to 644 on disk.0 it still failed with the same error.
>
> More help is needed.
>
>
> Thanks,
> Robert
>
>
> --
> Robert Schweikert                           MAY THE SOURCE BE WITH YOU
> SUSE-IBM Software Integration Center                   LINUX
> Tech Lead
> rjschwei at suse.com
> rschweik at ca.ibm.com
> 781-464-8147

--
Jaime Melis
Project Engineer
OpenNebula - The Open Source Toolkit for Cloud Computing
www.OpenNebula.org | jmelis at opennebula.org