On Wed, Nov 25, 2015, at 12:13, Simone Tiraboschi wrote:
On Mon, Nov 23, 2015 at 10:10 PM, Giuseppe Ragusa <giuseppe.ragusa(a)hotmail.com>
wrote:
> Hi all,
> I go on with my wishlist, derived from both solitary mumblings and community talks at
the the first Italian oVirt Meetup.
>
> I offer to help in coding (work/family schedules permitting) but keep in mind that
I'm a sysadmin with mainly C and bash-scripting skills (but hoping to improve my
less-than-newbie Python too...)
>
> I've sent separate wishlist messages for oVirt Node, oVirt Engine and VDSM.
>
> oVirt Self Hosted Engine Setup:
>
> *) allow virtual hardware customizations for locally-created Engine vm, specifically:
allow to add an arbitrary number of NICs (asking for MAC address and local bridge to
connect to) and maybe also an arbitrary number of disks (asking for size) as these seem to
be the only/most_useful items missing; maybe the prebuilt appliance image too may be
inspected by setup to detect a customized one and connect any further NICs to custom local
bridges (which the user should be asked for)
For 3.6.1 (it should be in 3.6.0 but it's bugged) you will be able to edit some
parameter of the engine VM from the engine (than of course you need to reboot to make them
effective).
I'm not sure if it's worth to make the setup more complex or if it's better
to keep it simple (single nic, single disk) and then let you edit the VM only from the
engine as for other VMs.
Thanks Simone for your reply!
You are right: I was bothering you with this setup wishlist item *mainly* because further
Engine vm modification was impossible/awkward/difficult before 3.6.1
Nonetheless I have seen many cases in which at least a second NIC would be absolutely
needed to complete the Engine installation: it is a well known best practice to keep the
management network (maybe conflated with the IPMI network in smaller cases) completely
isolated from other services and to allow only limited access to/from it, and that network
would be the ovirtmgmt-bridge-connected network (the only one available to the Engine, as
of now); now think of a kickstart-based Engine OS installation/update from a local
repository/mirror which would be reachable on a different network only (further access to
the User/Administration Web portal could have similar needs but could be more easily
covered by successive Engine vm modifications)
The "additional disks" part was (maybe "artificially") added by me out
of fantasy, but I know of at least one enterprise customer that by policy mandates
separate disks for OS and data (mainly on FC LUNs, to be honest, but FC is supported by
hosted Engine now, isn't it?)
I absolutely don't know how the setup code is structured (and the recent logical
"duplication" between mixins.py and vm.conf.in scares me a bit, actually ;), but
I naively hope that changing the two single hardcoded nic/hdd questions into two loops of
minimum 1 iteration (with a corresponding generalization of related otopi parameters)
should not increase the complexity too much (and could be an excuse to rationalize/unify
it further).
Obviously I could stand instantly corrected by anyone who really knows the code, but in
exchange I would gain for free some interesting pointers/insights into the setup
code/structure ;)
> Regards,
> Giuseppe
> _______________________________________________
> Users mailing list
> Users(a)ovirt.org
>
http://lists.ovirt.org/mailman/listinfo/users