19 Aug
2009
19 Aug
'09
12:15 p.m.
Hi Jocke,
On Wed, Aug 19, 2009 at 12:15:53PM +0100, Jocke Selin wrote:
On 15 Aug 2009, at 10:26, Andy Smith wrote:
On a normal machine with grub, a reboot gets you to the point where
the BIOS runs grub and grub gets kernel, initrd, etc.
On a Xen domU with pygrub, on domain creation pygrub runs in dom0
(the host server) to read your filesystem, extract your grub config
file, parse it to find the kernel, initrd etc., extract *those* to
a temporary directory and then create a domU with those files.
But on a reboot, Xen never actually destroys the virtual machine, it
just starts it again. That means it will have the same kernel
(pygrub is never run), same amount of RAM, most probably the same
size disk devices as well (I have never tested that bit).
Don't forget that you need to shut down and
boot again from xen
shell in order to get a new kernel; just rebooting uses the same
kernel again.
Why is this? I thought that, at least on Ubuntu a 'shutdown -r' picks up
the latest kernel?! Also, I'm having this issue on two of the VPS'
running Ubuntu at the
moment:
[...]
Get:13 http://apt-cacher.lon.bitfolk.com
hardy-updates/universe linux-
image-2.6.24-24-xen 2.6.24-24.59 [18.8MB]
45% [13 linux-image-2.6.24-24-xen 8967518/18.8MB 47%]
----
Both of them are stuck around 45% download...
The apt-cacher will proxy a connection to the real Ubuntu mirrors if
it doesn't have the file locally, so it could be that the mirrors
are overloaded.
Is it still happening?
Can you try other mirrors (change the part of the URL that is the
mirror hostname)?
Cheers,
Andy
--
http://bitfolk.com/ -- No-nonsense VPS hosting
"Reinventing the wheel is bad not only because it wastes time, but because
reinvented wheels are often square." -- Henry Spencer