SCIENTIFIC-LINUX-USERS Archives

February 2012

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Steven Haigh <[log in to unmask]>
Reply To:
Steven Haigh <[log in to unmask]>
Date:
Mon, 20 Feb 2012 12:30:21 +1100
Content-Type:
multipart/signed
Parts/Attachments:
text/plain (2837 bytes) , smime.p7s (4 kB)
On 20/02/2012 1:14 AM, Nico Kadel-Garcia wrote:
> On Sun, Feb 19, 2012 at 8:22 AM, Steven Haigh <[log in to unmask]
> <mailto:[log in to unmask]>> wrote:
>
>     I didn't agree with TUV in dropping Xen - however I'm pretty sure it
>     will return in EL7.
>
> I'm not sure about this. There are things I dislike intensely about KVM,
> specifically the lack of network configuration tools for setting up the
> requisite bridged networking and the resulting configuration chaos,
> especially if you want pair-bonded network ports. If I see one more such
> configuration written by a scripting monkey boy who *never tried
> disconnecting each of the cables to be sure it worked*, I'm going to be
> an unhappy camper.

Yeah - networking is a bit of an issue - but I've always set this up in 
the OS. Every setup is usually different and I don't believe a one size 
fits all approach ever works for something like this. I have 2 x VLANs 
on my Xen Dom0 server at home, and that goes into two bridges - one for 
each VLAN. Try getting a script that understands that!

> I'm also looking at the speed comparisons of KVM and Xen published at
> http://blog.xen.org/index.php/2011/11/29/baremetal-vs-xen-vs-kvm-redux/
> and followup at
> http://www.phoronix.com/scan.php?page=news_item&px=MTAyMTk
> <http://www.phoronix.com/scan.php?page=news_item&px=MTAyMTk>. And
> overall, the performances between the latest manually patched kernels
> with Xen and the latest KVM are..... hardly significant for typical use.
> So is it worth the extra sweat and tears to maintain the distinct
> architecture.
> Do you really see enough benefit to justify the work? And what makes you
> think that our favorite upstream vendor will revisit Xen support for
> release 7?

I don't think it'll be that much work. Xen is still maintained for 
Fedora - so its just a matter of rebuilding the package for EL7. As EL7 
will probably be based on a 3.x kernel, all the inner workings to run as 
a Xen Dom0 will be available in the kernel - they just need to be 
enabled. Kernel 2.6.32 that EL6 is built with still needs patches.

As for is it worth the effort? I believe so. I have multiple Xen systems 
and migrating them will be a pain in the backside. I've fine tuned the 
config over the years and its not worth starting again to maintain things.

As for me personally, I'd already be building these packages for myself 
- even if I didn't release them - so the extra work for me is just the 
32 bit packages - which I am not easily able to test. Thankfully, after 
setting up a dev environment with mock, its relatively easy to manage it 
all.

I figure for the thousands of downloads the 64bit packages have had, 
there are many people in the same boat.

-- 
Steven Haigh

Email: [log in to unmask]
Web: http://www.crc.id.au
Phone: (03) 9001 6090 - 0412 935 897
Fax: (03) 8338 0299



ATOM RSS1 RSS2