> On 28 Apr 2018, at 13:37, Mark Raynsford
> <list+org.freebsd.virtualizat...@io7m.com> wrote:
>
> On 2018-04-28T09:08:42 +0300
> Daniel Braniss <da...@cs.huji.ac.il> wrote:
>
>> since the clients and the server are sharing the zfs volume,
>> I’m doing the following:
>> on the server I did:
>> zfs create -sV 4G h/root.ro <http://root.ro/>
>> newfs /dev/zvol/h/root.ro <http://root.ro/>
>> mount /dev/zol/h/root.ro <http://root.ro/> /mnt
>> copy a working root image to it.
>> umount /mnt
>> the clients then mount it as ro,
>> the vm conflg file has:
>> disk0_type=virtio-blk”
>> disk0_name=“/dev/zvol/h/root.ro <http://root.ro/>”
>> disk0_dev=“custom”
>>
>> one solution to the fact that the root is read-only is to use unionfs
>> (probably nullfs will do too)
>>
>> the only problem I have is updating the image.
>
> Wow, didn't know this was possible. Is this safe? Two essentially
> independent operating system instances being able to write to the same
> zvol?
that’s why it’s mounted rear-only in the client!
each client can get another vol for writing, ie /var
if you want to have ‘permanent’ changes that will survive reboots.
updating on the server is possible, but
1- the changes might not be seen by the client
2- opened files will have issues
btw, point 2 is also true for NFS.
danny
>
> --
> Mark Raynsford | http://www.io7m.com
>
_______________________________________________
freebsd-virtualization@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-virtualization
To unsubscribe, send any mail to
"freebsd-virtualization-unsubscr...@freebsd.org"
