[rhos-list] Problem with Snapshots.

[Perry Myers]

On 04/10/2013 12:04 PM, Minton, Rich wrote:
> I’m having an issue when I create an instance from a Snapshot image…
> 
>  
> 
> The instance is created from the snapshot fine but in the console log I
> see errors when trying to get to the metadata service, host is unreachable.
> 
>  
> 
> 2013-04-10 11:53:38,716 - url_helper.py[WARNING]: Calling
> 'http://169.254.169.254/2009-04-04/meta-data/instance-id' failed
> [0/120s]: url error [[Errno 101] Network is unreachable]
> 
>  
> 
> I also get an error regarding eth0:
> 
> Bringing up interface eth0:  Device eth0 does not seem to be present,
> delaying initialization. [FAILED]
> 
>  
> 
> Cloud-init starts ok and I see another entry regarding networking:
> 
>  
> 
> ci-info: +++++++++++++++++++++++Net device info++++++++++++++++++++++++
> 
> ci-info: +--------+-------+-----------+-----------+-------------------+
> 
> ci-info: | Device |   Up  |  Address  |    Mask   |     Hw-Address    |
> 
> ci-info: +--------+-------+-----------+-----------+-------------------+
> 
> ci-info: |   lo   |  True | 127.0.0.1 | 255.0.0.0 |         .         |
> 
> ci-info: |  eth1  | False |     .     |     .     | fa:16:3e:6f:24:24 |
> 
> ci-info: +--------+-------+-----------+-----------+-------------------+
> 
> ci-info: !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!Route info
> failed!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
> 
>  
> 
> It looks like there is an eth1 but no eth0.  Looking further I found two
> entries in /etc/udev/rules.d/70-persisten-net.rules for eth0 and eth1.
> If I delete both lines and reboot the interfaces are created properly
> and the the instance can reach the metadata service and all works as it
> should.
> 
>  
> 
> My question is, is there a way to keep the second NIC from being created
> and have eth0 get the new hardware address? Should I delete the
> 70-persisten-net.rules file before creating the snapshot or is there an
> easier way?

Did you 'sysprep' the guest image before importing it into glance?

The udev rule you mention above needs to be removed/disabled from guest
images in order to prevent this I think.

There is a tool called virt-sysprep that you can run on guest images
prior to import and one of the default things it does is to remove that
troublesome udev rule

I've added the libguestfs maintainer (which virt-sysprep is part of) to
provide any add'l insight here.

SteveG/ChrisN, do you guys have usage of virt-sysprep for RHOS images
included in either a kbase or the formal docs?

Cheers,

Perry