[CRIU] Using p.haul migration failure

Pavel Emelyanov xemul at parallels.com
Wed Nov 5 03:10:30 PST 2014


On 11/02/2014 06:08 AM, Sowmini Varadhan wrote:
> On (10/28/14 15:57), Pavel Emelyanov wrote:
>>
>> Please, note, that your dump stage will fail anyway, as there are some
>> mount points in your program, that CRIU doesn't support for now. Let's
>> check what it is.
>>
>> Thanks,
>> Pavel
> 
> yes, I must be hitting this point. 
> 
> having hacked the python to work around the veth.pair issue,
> I'm now able to get further, but now I hit this error
> 
>     Final dump and restore
>             Making directory /var/local/p.haul-fs/dmp-sYVUup-14.11.01-21.43/img/3
>             Issuing dump command to service
>                     Downing veth6GI2X1
>                     Notify (network-lock)
>                     Upping veth6GI2X1
>                     Notify (network-unlock)
>     Traceback (most recent call last):
>       File "/home/sowmini/src/p.haul/p.haul", line 39, in <module>
>         worker.start_migration()
>       File "/home/sowmini/src/p.haul/p_haul_iters.py", line 182, in start_migration
>         raise Exception("Dump failed")
>     Exception: Dump failed
> 
> Examining the log file, seems like the issue is this
> 
>     (00.054526) Error (files-reg.c:792): Unable to look up the 18 mount
>     (00.054542) BUF 0x7f3998ccc000 >
>     (00.054549) ----------------------------------------
>     (00.054552) Error (cr-dump.c:1596): Dump files (pid: 2681) failed with -1

There are problems with mount points. Can you do "criu dump -R" on your container
successfully?

> full log file (criu_dump.3.log) is attached.
> Where is this now getting stuck? Any workarounds?
> 
> --Sowmini
> 



More information about the CRIU mailing list