[Planetlab-devel] pl/xen

[David Irwin]

Hi Marc,

Sorry for the delayed reply.  There seem to be many deadlines this
month.  There are answers to some of your questions below.

> Our bootcd kernel also happens to be our PL production kernel.  Are the
> kexec kernel changes very intrusive?

I believe they are intrusive and I haven't tried applying the patch to
a PL Xen kernel.  I replaced the kernel used by the boot CD with a
xen-unstable kernel with the patches applied.  There seems to be quite
a bit of movement on kexec on the xen-devel mailing lists, but I'm not
sure when the changes will go into Xen.  There was a post that kexec
support was in Xen 3.0.4, but I think it was referring to dom0 kexec
although I'm not quite sure (I appended the Xen 3.0.4 announcment
below).

> WRT managing software updates correctly, xen-based production PL
> systems simply would need to be added to a node group whose yum
> repository would have xen PLKs rather than regular PLKs.  I think that
> is about it!

That's great.  The disk image on the boot cd would also need the
libraries/executable for invoking the "xc_kexec" utility (Xen's
version of the user-space kexec utility).  You would also need to
modify your chain boot script to use this utility for Xen PL nodes.
When I deploy a MyPLC I create a new boot cd image with the necessary
utilities and the modified scripts.

> Please remind me with which PL provided kernel you were working.  We've
> run into some 2.6.18 / FC5 badness that prompted me to revert back to a
> 2.6.17-based FC5 kernel, but I have not checked on its Xen capabilities.

I did not end up using your PL/Xen kernel.  It had some things that
were incompatible with the most recent version of Xen (I sent a
previous email related to this).  I checked out a recent (at the time)
PL kernel from the CVS tree and compiled it myself, and it didn't
exhibit this problem.  This is a 2.6.18 meaning that the 2.6.17 would
need to be tested.

The PL Fedora distribution also seems to have the problem discussed in
this thread:  http://www.redhat.com/archives/fedora-xen/2006-October/msg00158.html
(e.g., errors in the form of "4gb seg fixup, process named (pid 1266),
cs:ip 73:00561378 printk: 484562 messages suppressed.")  I posted
previously about this:  it is a well-known issue with Xen (and
Fedora).  This needs to be fixed for production.

-David







Message: 3
Date: Wed, 20 Dec 2006 16:25:14 +0000
From: Keir Fraser <keir at xensource.com>
Subject: [Xen-users] FW: Xen 3.0.4 released!
To: Xen devel list <xen-devel at lists.xensource.com>,
       "xen-users at lists.xensource.com" <xen-users at lists.xensource.com>
Message-ID: <C1AF146A.6A73%keir at xensource.com>
Content-Type: text/plain;       charset="US-ASCII"

Folks,

We're pleased to announce the official release of xen 3.0.4!

This is largely an opportunistic stabilising release for HVM guests, due to
the large amount of work in that area of the code since 3.0.3. These
enhancements have in particular improved support for SMP and ACPI Linux and
Windows operating systems.

Other highlights of this release include:
 - support for kexec/kdump of Xen and domain 0;
 - graphical framebuffer support for paravirtualised guests;
 - preview support for the new XenAPI management interfaces;
 - enhanced support for IA64 (IPF) and Power systems.

Since 3.0.4 is an interim release, certain features such as HVM save/restore
will now be part of Xen 3.0.5 which we expect to release in early 2007.

You can get the source using mercurial from:
 http://xenbits.xensource.com/xen-3.0.4-testing.hg

Source and binary tarballs, and RPMs, will be made available from:
 http://www.xensource.com/downloads

Cheers,
Keir (on behalf of the whole Xen dev team)