Re: [Xen-users] Live checkpointing not working in 3.4.x?

[Pasi Kärkkäinen <pasik@xxxxxx>]

On Fri, Mar 05, 2010 at 02:04:19AM -0600, Tom Verbiscer wrote:
> Just tried it on 2.6.18-164.11.1.el5xen and it has the same issue.
>
> As far as using the default RHEL5 stuff, CentOS at least only comes with  
> Xen 3.0.3 which doesn't support live checkpointing.
>

RHEL5/CentOS5 has Xen *hypervisor* version 3.1.2 + a lot of patches from Redhat.
You can verify that with "xm info" - see the xen_major,xen_minor,xen_extra 
fields.

Just the xen management tools are based on 3.0.3 in EL5.

> I tried it with Xen 3.4.0 and a Fedora 9 domU just to try it out with a  
> somewhat recent domU kernel.  When I did that, the command ('xm save -c  
> <domain#> <file>') hung.  The VM never suspended, it just kept running. 
>
> Any ideas?
>

Maybe try with the latest http://xenbits.xen.org/xen-3.4-testing.hg (3.4.3-rc3).

-- Pasi

> Thanks much,
> Tom
>
>
>
> Pasi Kärkkäinen wrote:
>> On Thu, Mar 04, 2010 at 08:05:24AM -0600, Tom Verbiscer wrote:
>>   
>>> Normal 'xm save' and 'xm restore' works just fine.  My PV guest kernel is:
>>>
>>>     
>>
>> Ok.
>>
>>   
>>> [root@ovm-pv-01 ~]# uname -a
>>> Linux ovm-pv-01.example.com 2.6.18-164.el5xen #1 SMP Thu Sep 3 
>>> 04:41:04  EDT 2009 x86_64 x86_64 x86_64 GNU/Linux
>>>
>>>     
>>
>> This kernel should be OK. You should update to the latest hotfix release 
>> though (-164.something)
>>
>> Does it work with the default Xen 3.1.2 that comes with RHEL5? 
>>
>> -- Pasi
>>
>>   
>>> Thanks,
>>> Tom
>>>
>>>
>>> Pasi Kärkkäinen wrote:
>>>     
>>>> On Thu, Mar 04, 2010 at 12:12:56AM -0600, Tom Verbiscer wrote:
>>>>         
>>>>> I've been banging my head against a wall for a couple days now.  
>>>>> Does  anyone know if live checkpointing ('xm save -c') is 
>>>>> currently working in  3.4.x?  I've now tried with 3.4.0 on 
>>>>> OracleVM, 3.4.1 on CentOS 5.4 and  3.4.2 on OpenSolaris.  Each 
>>>>> platform gives me the same results.  It  seems like the suspend 
>>>>> works but does not release the devices so when  the resume runs, 
>>>>> it freaks because the devices are already attached.  I  don't 
>>>>> know enough about Xen to know if the devices are supposed to  
>>>>> remain attached (because it doesn't destroy the domain) or not.  
>>>>> Every  time I try to live checkpoint the VM winds up suspended 
>>>>> and the only way  to bring it back to life is to run 'xm destroy' 
>>>>> on it and then 'xm  resume'.  I'll be happy to provide more logs 
>>>>> if I've leaving something  out.  The following is on a OracleVM 
>>>>> hypervisor (yes, OracleVM doesn't  support checkpointing but the 
>>>>> results are the same with vanilla Xen).   Also doesn't matter if 
>>>>> I use a file backend device for the disk or a   physical device 
>>>>> or a file on an NFS share, same result.
>>>>>
>>>>>             
>>>> does normal "xm save" and then "xm restore" work for you? 
>>>>
>>>> What's the guest kernel version? save/restore heavily depends on 
>>>> the  guest kernel version/features (for pv guests).
>>>>
>>>> -- Pasi
>>>>
>>>>         
>>>>> Thanks,
>>>>> Tom
>>>>>
>>>>> [root@compute-01 ~]# rpm -qa | grep xen
>>>>> xen-devel-3.4.0-0.0.23.el5
>>>>> xen-tools-3.4.0-0.0.23.el5
>>>>> xen-debugger-3.4.0-0.0.23.el5
>>>>> xen-3.4.0-0.0.23.el5
>>>>> xen-64-3.4.0-0.0.23.el5
>>>>> [root@compute-01 ~]# uname -a
>>>>> Linux compute-01.example.com 2.6.18-128.2.1.4.9.el5xen #1 SMP Fri 
>>>>> Oct 9  14:57:31 EDT 2009 i686 i686 i386 GNU/Linux
>>>>>
>>>>> [root@compute-01 ~]# cat /OVS/running_pool/1_ovm_pv_01_example_com/vm.cfg
>>>>> bootargs = 'bridge=xenbr0,mac=00:16:3E:AA:EB:08,type=netfront'
>>>>> bootloader = '/usr/bin/pypxeboot'
>>>>> disk = ['file:/tmp/System.img,xvda,w']
>>>>> maxmem = 512
>>>>> memory = 512
>>>>> name = '1_ovm_pv_01_example_com'
>>>>> on_crash = 'restart'
>>>>> on_reboot = 'restart'
>>>>> uuid = '7408c627-3232-4c1d-b5e3-1cf05cb015c8'
>>>>> vcpus = 1
>>>>> vfb = ['type=vnc,vncunused=1,vnclisten=0.0.0.0,vncpasswd=<removed>']
>>>>> vif = ['bridge=xenbr0,mac=00:16:3E:AA:EB:08,type=netfront']
>>>>> vif_other_config = []
>>>>>
>>>>>             
>>>> _______________________________________________
>>>> Xen-users mailing list
>>>> Xen-users@xxxxxxxxxxxxxxxxxxx
>>>> http://lists.xensource.com/xen-users
>>>>         
>>
>> _______________________________________________
>> Xen-users mailing list
>> Xen-users@xxxxxxxxxxxxxxxxxxx
>> http://lists.xensource.com/xen-users
>>   

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users