Fix corruption issue with packer build

[sc68cal]

I was having the same issues as #13 - where the script files in /mnt/flash were hosed and the management interface wouldn't come up. But instead of doing base64 encoding, I looked into the documentation for Packer to find out how to stop Packer from killing off the VM without syncing the filesystem and flushing the buffers.

Arista EOS doesn't have a way via FastCli to halt the OS, however you can jump over into bash and then use the standard linux shutdown utility to get the VM to halt.

[sc68cal]

I also added the eXecute bit to the initialize_ma1.sh script so that more recent vEOS images that configure the following, will work correctly.

event-handler dhclient
   trigger on-boot
   action bash sudo /mnt/flash/initialize_ma1.sh
!

[sc68cal]

I think also more recent VMDK's of vEOS-lab are packaging their own initialize_ma1.sh file - so we may not even need to keep that script in this git repo