On Fri, 15 Apr 2005, David Mohring wrote:
a) Whats the chances of a stable ( enough ) Xen making it into the
final
release of FC4?
Pretty big, the current xen seems to be reasonably stable.
That doesn't mean all aspects will be easy to use, though...
b) Is it going to be safe/stable/secure for the domain U kernels use
LVM
mounts? Or are loopback file systems, mounted by the domain 0 kernel,
the only effective choice?
I've been using LVM volumes for Xen guests for a while now
and it's working fine.
c) Is it worth allocating swap partitions for domain U systems? Is
it
safe/possible to use LVM for this?
Yes, LVM works great.
d) Is the combination of X-server/AGP/sound going to available and
stable on domain 0 kernels?
It already is, on some hardware ;)
If it isn't working right on your hardware, please consider
filing a bug on
http://bugzilla.xensource.com/ (which is where
most Xen development takes place).
e) Will the actual performance of the non-test final release kernels
approach the performance of the recent Xen demos?
It should.
Have the Fedora Core core developer given any consideration to a
single
"xen" init.d script which would :
1) Setup the virtual networking for the domain 0 and domain U hosted
systems. Virtual bridging + remote DHCP/static addressing
OR internal DHCPd + NAT + plus forwarding
2) Allocate file system mount rights on a first come first served basis,
with fall over of NFS mounting if the first serves the requested
partition.
When it comes to things like this, I try to not do anything
Fedora specific, but instead develop things together with the
Xen community.
If you would like to get involved with the development of
such scripts, you may want to subscribe to the xen-devel
mailing list on
lists.xensource.com.
There should be an easier way to configuring the combinations of
Fedora,Grub,LVM and Xen than manually editing configure files and CLI
launching instances.
Lastly. Any word on getting a version of Anaconda to work under
Domain U
Server level? Ie VNCserver mode with the domain 0 host granting lvm
partition/group access.
We've got plans for fixing these issues, but I don't know
when we'll get around to them. Helping with the Xen 3.0
release is a higher priority at the moment.
--
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan