[one-users] Save As and Detach option

Tue Nov 27 05:30:31 PST 2012

Images are only saved when you "power off" a machine with SHUTDOWN or
CANCEL but not with DELETE. Delete command is ment to get rid of a VM
that is in a wrong state or we just don't care about so no image
saving is done.

>From what I see in the logs you are executing delete. If you just want
to unplug the VM (not an ordered shutdown) you can use cancel. This
will trigger the correct epilog that saves the images.

On Tue, Nov 27, 2012 at 2:25 PM, Roberto Rosende Dopazo
<rrosende at cesga.es> wrote:
> Hi
>
>
> El 27/11/12 13:58, Ruben S. Montero escribió:
>
> Hi,
>
> Just to add a few considerations:
>
> 1.- Save_as images are not created on the datastore, just an ID is generated
> as a placeholder for the new image (this is the fstype: save_as). So, it is
> normal that no image (actual path) is created after a save_as operation, and
> the state is supposed to be locked (waiting for the real image) as reported.
> I do not see any issue here.
>
> Ok, I understood that file was created at that moment.
>
>
> 2.- The save_as image is moved to the final destination when the VM is
> shutdown, shutdown means onevm  shutdown (delete or stop  will not move the
> image, for example). Could you send the output vm.log of the image after the
> shutdown. You should look for messages about a  MVDS  command.
>
> The complete log for that machine, there is not MVDS message. From onevm
> show image was created at REGISTER TIME  : 11/27 11:36:26
>
> [root at cloud one]# cat 138.log
> Tue Nov 27 11:34:56 2012 [DiM][I]: New VM state is ACTIVE.
> Tue Nov 27 11:34:56 2012 [LCM][I]: New VM state is PROLOG.
> Tue Nov 27 11:34:56 2012 [VM][I]: Virtual Machine has no context
> Tue Nov 27 11:34:58 2012 [TM][I]: clone: Cloning
> /var/lib/one/datastores/100/3dfef6d01e7e624f13ef13698ed04710 in
> nubacesga-01-4:/var/lib/one/datastores/0/138/disk.0
> Tue Nov 27 11:34:58 2012 [TM][I]: ExitCode: 0
> Tue Nov 27 11:35:18 2012 [TM][I]: mkimage: Making filesystem of 1024M and
> type swap at nubacesga-01-4:/var/lib/one/datastores/0/138/disk.1
> Tue Nov 27 11:35:18 2012 [TM][I]: ExitCode: 0
> Tue Nov 27 11:35:18 2012 [LCM][I]: New VM state is BOOT
> Tue Nov 27 11:35:18 2012 [VMM][I]: Generating deployment file:
> /var/lib/one/138/deployment.0
> Tue Nov 27 11:35:18 2012 [VMM][I]: ExitCode: 0
> Tue Nov 27 11:35:18 2012 [VMM][I]: Successfully execute network driver
> operation: pre.
> Tue Nov 27 11:35:19 2012 [VMM][I]: ExitCode: 0
> Tue Nov 27 11:35:19 2012 [VMM][I]: Successfully execute virtualization
> driver operation: deploy.
> Tue Nov 27 11:35:19 2012 [VMM][I]: ExitCode: 0
> Tue Nov 27 11:35:19 2012 [VMM][I]: Successfully execute network driver
> operation: post.
> Tue Nov 27 11:35:19 2012 [LCM][I]: New VM state is RUNNING
> Tue Nov 27 11:38:01 2012 [LCM][I]: New VM state is CLEANUP.
> Tue Nov 27 11:38:01 2012 [DiM][I]: New VM state is DONE
> Tue Nov 27 11:38:01 2012 [VMM][W]: Ignored: LOG I 138 Driver command for 138
> cancelled
> Tue Nov 27 11:38:01 2012 [TM][W]: Ignored: LOG I 138 ExitCode: 0
> Tue Nov 27 11:38:01 2012 [TM][W]: Ignored: LOG I 138 ExitCode: 0
> Tue Nov 27 11:38:01 2012 [TM][W]: Ignored: LOG I 138 delete: Deleting
> /var/lib/one/datastores/0/138
> Tue Nov 27 11:38:01 2012 [TM][W]: Ignored: LOG I 138 ExitCode: 0
> Tue Nov 27 11:38:01 2012 [TM][W]: Ignored: TRANSFER SUCCESS 138 -
> Tue Nov 27 11:38:01 2012 [VMM][W]: Ignored: LOG I 138 ExitCode: 0
> Tue Nov 27 11:38:01 2012 [VMM][W]: Ignored: LOG I 138 Successfully execute
> virtualization driver operation: cancel.
> Tue Nov 27 11:38:01 2012 [VMM][W]: Ignored: LOG I 138 ExitCode: 0
> Tue Nov 27 11:38:01 2012 [VMM][W]: Ignored: LOG I 138 Successfully execute
> network driver operation: clean.
> Tue Nov 27 11:38:01 2012 [VMM][W]: Ignored: CANCEL SUCCESS 138 -
>
>
> Cheers
>
> Ruben
>
>
> On Tue, Nov 27, 2012 at 12:41 PM, Roberto Rosende Dopazo <rrosende at cesga.es>
> wrote:
>>
>> Hi,
>>
>> El 27/11/12 11:50, Javier Fontan escribió:
>>
>>> Hi,
>>>
>>> I'm struggling to understand the situation you are describing. Correct
>>> me if I am wrong.
>>>
>>> You mention dom0. Are you using Xen? If that is true disregard what I
>>> said about changing dynamic_ownership parameter in libvirt as Xen
>>> drivers do not use libvirt for hypervisor communication.
>>
>> It's kvm image and dynamic_ownership was wrong because I misunderstood and
>> put it to 1, now is 0 and owner and group is not changed.
>> But option save as still doesn't work, I tried it directly from frontend
>> shell instead SunStone and I got that:
>>
>>
>> [grid at cloud ~]$ onevm show 138
>> VIRTUAL MACHINE 138 INFORMATION
>> ID                  : 138
>> NAME                : one-138
>> USER                : emi
>> GROUP               : oneadmin
>> STATE               : ACTIVE
>> LCM_STATE           : RUNNING
>> RESCHED             : No
>> HOST                : nubacesga-01-4
>> START TIME          : 11/27 11:34:32
>> END TIME            : -
>> DEPLOY ID           : one-138
>>
>> VIRTUAL MACHINE MONITORING
>> USED MEMORY         : 0K
>> NET_RX              : 0K
>> USED CPU            : 0
>> NET_TX              : 0K
>>
>> PERMISSIONS
>> OWNER               : um-
>> GROUP               : ---
>> OTHER               : ---
>>
>> VIRTUAL MACHINE TEMPLATE
>> CPU="1"
>> DISK=[
>>   BUS="virtio",
>>   CLONE="YES",
>>   CLUSTER_ID="100",
>>   DATASTORE="GRID EMC",
>>   DATASTORE_ID="100",
>>   DEV_PREFIX="hd",
>>   DISK_ID="0",
>>   DRIVER="qcow2",
>>   IMAGE="GRID SL6_emi_cream_orig",
>>   IMAGE_ID="11",
>>   READONLY="NO",
>>   SAVE="NO",
>> SOURCE="/var/lib/one/datastores/100/3dfef6d01e7e624f13ef13698ed04710",
>>   TARGET="hda",
>>   TM_MAD="shared",
>>   TYPE="FILE" ]
>> DISK=[
>>   DEV_PREFIX="hd",
>>   DISK_ID="1",
>>   SIZE="1024",
>>   TARGET="hdb",
>>   TYPE="swap" ]
>> GRAPHICS=[
>>   KEYMAP="es",
>>   LISTEN="0.0.0.0",
>>   PORT="6038",
>>   TYPE="vnc" ]
>> MEMORY="2512"
>> NAME="one-138"
>> NIC=[
>>   BRIDGE="virbrPUBLIC",
>>   CLUSTER_ID="100",
>>   IP="193.144.35.36",
>>   MAC="02:00:c1:90:23:24",
>>   MODEL="e1000",
>>   NETWORK="red-193.144.35",
>>   NETWORK_ID="0",
>>   VLAN="NO" ]
>> OS=[
>>   ROOT="hda1" ]
>> PRODUCT="EMI-test"
>> REQUIREMENTS="CLUSTER_ID = 100"
>> TEMPLATE_ID="6"
>> VCPU="1"
>> VMID="138"
>>
>> VIRTUAL MACHINE HISTORY
>>  SEQ HOST            REASON           START            TIME PROLOG_TIME
>>    0 nubacesga-01-4  none    11/27 11:34:56    0d 00h01m09s    0d
>> 00h00m22s
>>
>>
>>
>>
>> [grid at cloud ~]$ onevm saveas 138 0 changes01
>> Image ID: 31
>>
>>
>>
>> [grid at cloud ~]$ oneimage list
>>   ID USER       GROUP      NAME DATASTORE     SIZE TYPE PER STAT RVMS
>> [...]
>>   31 grid       oneadmin   changes01 GRID EMC      2.2G OS    No lock    0
>>
>>
>> [grid at cloud ~]$ oneimage show 31
>> IMAGE 31 INFORMATION
>> ID             : 31
>> NAME           : changes01
>> USER           : grid
>> GROUP          : oneadmin
>> DATASTORE      : GRID EMC
>> TYPE           : OS
>> REGISTER TIME  : 11/27 11:36:26
>> PERSISTENT     : No
>> SOURCE         :
>> /var/lib/one/datastores/100/08cee146d8bff43c6c61794f1ba947d3
>> FSTYPE         : save_as
>> SIZE           : 2.2G
>> STATE          : lock
>> RUNNING_VMS    : 0
>>
>> PERMISSIONS
>> OWNER          : um-
>> GROUP          : ---
>> OTHER          : ---
>>
>> IMAGE TEMPLATE
>> DEV_PREFIX="hd"
>> SAVED_DISK_ID="0"
>> SAVED_IMAGE_ID="11"
>> SAVED_VM_ID="138"
>> SAVE_AS="YES"
>>
>>
>>
>> From onevm help I understood that now it's supposed I have the file :
>>
>> saveas <vmid> <diskid> <img_name>
>>         Sets the specified VM disk to be saved in a new Image. The Image
>> is
>>         created immediately, but the contents are saved only if the VM is
>>         shut down gracefully (i.e., using 'onevm shutdown' and not
>>         'onevm delete')
>>
>>
>> But it does not exist:
>> [grid at cloud ~]$ file
>> /var/lib/one/datastores/100/08cee146d8bff43c6c61794f1ba947d3
>> /var/lib/one/datastores/100/08cee146d8bff43c6c61794f1ba947d3: cannot open
>> `/var/lib/one/datastores/100/08cee146d8bff43c6c61794f1ba947d3' (No such file
>> or directory)
>>
>> After a shutdown file still does not exist but image is marked as READY
>>
>> [grid at cloud ~]$ oneimage list
>>   ID USER       GROUP      NAME DATASTORE     SIZE TYPE PER STAT RVMS
>>    [...]
>>   31 grid       oneadmin   changes01 GRID EMC      2.2G OS    No rdy     0
>>
>>
>> And file doesn't exists anyway, from oned.log
>> 492501-Tue Nov 27 11:36:26 2012 [ReM][D]: Req:5808 UID:0 AclInfo invoked
>> 492502-Tue Nov 27 11:36:26 2012 [ReM][D]: Req:5808 UID:0 AclInfo result
>> SUCCESS, "<ACL_POOL><ACL><ID>0..."
>> 492503-Tue Nov 27 11:36:26 2012 [ReM][D]: Req:4608 UID:2
>> VirtualMachineSaveDisk invoked, 138, 0, "changes01", ""
>> 492504-Tue Nov 27 11:36:26 2012 [ImM][I]: Creating disk at  of 2216Mb
>> (type: save_as)
>> 492505-Tue Nov 27 11:36:26 2012 [ReM][D]: Req:4608 UID:2
>> VirtualMachineSaveDisk result SUCCESS, 31
>> 492506-Tue Nov 27 11:36:26 2012 [ImG][D]: Message received: LOG I 31
>> ExitCode: 0
>> 492507-
>> 492508-Tue Nov 27 11:36:26 2012 [ImM][I]: ExitCode: 0
>> 492509:Tue Nov 27 11:36:26 2012 [ImG][D]: Message received: MKFS SUCCESS
>> 31 /var/lib/one/datastores/100/08cee146d8bff43c6c61794f1ba947d3
>> 492510-
>>
>>
>> Cheers,
>>  Roberto
>>
>>
>>
>>
>>> First line I see in the in the log file seems seriously broken:
>>>
>>> 426622-Mon Nov 26 14:06:42 2012 [ImM][I]: Creating disk at  of 2216Mb
>>> (type: save_as)
>>>
>>> It's not saying where it is creating the image (source in OpenNebula
>>> parlance) and give an strange type. Could you please send us the VM
>>> and Image templates from onevm and oneimage show? Maybe I am missing
>>> something and other developers could help me here.
>>>
>>> I also see this line that says the image was created:
>>>
>>> 426627:Mon Nov 26 14:06:42 2012 [ImG][D]: Message received: MKFS
>>> SUCCESS 25 /var/lib/one/datastores/100/a3e076f377a855dc76f44ac373c9b0ca
>>>
>>> But you are telling me that the image does not exist. Are you checking
>>> the hypervisor node or the frontend?
>>>
>>> More comments inline
>>>
>>>
>>> On Mon, Nov 26, 2012 at 2:33 PM, Roberto Rosende Dopazo
>>> <rrosende at cesga.es> wrote:
>>>>
>>>> [...]
>>>> After delete machine image state is showed as READY in Sunstone.
>>>> In dom0 deployment dir and file are still present and disk files disk.0
>>>> and
>>>> disk.1 have their user and group changed to root.root and image file
>>>> doesn't
>>>> exist
>>>> A new file appears
>>>> # file /var/lib/one/datastores/0/123/checkpoint
>>>> /var/lib/one/datastores/0/123/checkpoint: data
>>>
>>> This is really strange. After EPILOG (the state that comes after
>>> successful shutdown, delete or cancel) the VM directory in the system
>>> datastore is deleted, so those files should not be there. Also, then
>>> only way OpeNebula tells Xen to create checkpoint is on SAVE, STOP or
>>> MIGRATE. Are you sure that you have deleted the VM? It looks to me
>>> that the image was suspended or stopped. Bear in mind that save_as is
>>> not done in that state.
>>>
>>>> Also with root.root but 666 file permission and in this case directory
>>>> and
>>>> files of this machine are never deleted from running datastore
>>>
>>> The non deletion of images is consistent my the SAVE/STOP theory.
>>>
>>> Cheers
>>>
>>> --
>>> Javier Fontán Muiños
>>> Project Engineer
>>> OpenNebula - The Open Source Toolkit for Data Center Virtualization
>>> www.OpenNebula.org | jfontan at opennebula.org | @OpenNebula
>>
>>
>> _______________________________________________
>> Users mailing list
>> Users at lists.opennebula.org
>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>
>
> --
> Ruben S. Montero, PhD
> Project co-Lead and Chief Architect
> OpenNebula - The Open Source Solution for Data Center Virtualization
> www.OpenNebula.org | rsmontero at opennebula.org | @OpenNebula
>
>

-- 
Javier Fontán Muiños
Project Engineer
OpenNebula - The Open Source Toolkit for Data Center Virtualization
www.OpenNebula.org | jfontan at opennebula.org | @OpenNebula