[elbe-devel] elbe initvm attach -> login does not work after upgrading from 1.1 to 1.2 stable

[Richard Cochran]

On Wed, Dec 13, 2017 at 06:21:29AM +0100, Manuel Traut wrote:
> On Tue, Dec 12, 2017 at 04:40:05PM +0000, Lukasz Walewski wrote:
> > However, the whole "initvm create" fails on the partitioning stage, which is probably unrelated to the selection of the VM type:
> > 
> > ...
> > Starting up the partitioner  ... 13%... 22%... 31%... 40%... 50%... 63%... 72%... 81%... 90%... 100%
> > Guided partitioning  ... 20%... 40%... 60%... 80%
> > Partitions formatting  ... 33%
> > Partition disks
> > ---------------
> > 
> > The attempt to mount a file system with type ext3 in Virtual disk 1, partition
> > #1 (vda1) at / failed.
> > 
> > You may resume partitioning from the partitioning menu.
> > Do you want to resume partitioning?
> >   1: Yes [*]  2: No
> > Prompt: '?' for help, default=1>
> > 
> > Any hints how to fix this?
> 
> The problem is in the debian installer, if you answer the question with No and
> go to a shell:
> 
> ~ # insmod /lib/modules/3.16.0-4-amd64/kernel/fs/ext4/ext4.ko
> insmod: ERROR: could not insert module /lib/modules/3.16.0-4-amd64/kernel/fs/ext4/ext4.ko: Unknown symbol in module
> 
> The installer is provided by the 'elbe-bootstrap' package. Probably Debian
> uses a new kernel in Debian 8.10 released on 9th of December. So the old
> Kernel provided by elbe-bootstrap is not able to load the 'newer' kernel modules
> downloaded from the configured primary mirror.

What was the resolution of this problem?

I have seen this exact problem again yesterday.  The context was MacOS
running Ubuntu 16.04 inside vmware (with nesting enabled).  Elbe ran
from git on current branch devel/elbe-2.0.  Command was this:

    elbe init --directory=~/elbe-vm --nesting examples/elbe-init-with-ssh.xml

and then 'make' in ~/elbe-vm.

The error does not seem to be related to running under vmware.
Strangely, when I ran the same under bare metal ubuntu 16.4 the error
did not appear.

Confused,
Richard