Re: [vserver] util-vserver & cpuset bug ?

From: Herbert Poetzl <>
Date: Wed 15 Jul 2009 - 23:03:55 BST
Message-ID: <>

On Wed, Jul 15, 2009 at 01:24:24PM +0100, Ed W wrote:
> Herbert Poetzl wrote:

> >sidenote: the 'enter' is not the best way to logon to
> >a guest anyways, it is more a backdoor, but it should
> >get as close as possible to a real logon, except for
> >environment and shell specifica

> What do you recommend instead?

well, you have a bunch of options, depending on the
specific usecase ... e.g.

 - use a real console (serial or vga)
 - use a virtual console
 - use a pipe or unix socket
 - use a network socket

> SSH is a given, but isn't always a straightforward option
> if IP addresses are constrained

AFAIK, there is no constraint on private IP addresses
(yet, although the IPV6 folks would like to see that
on IPV4 :), so the public IP address constrains are
not a real reason for not using ssh or telnet

> Did I miss another better way to login if you are already
> on the host?

maybe I should rephrase my original answer and state
here that 'vserver - enter' (for good reason not called
logon) is not really a logon to the guest, instead it
moves the current process into the various namespaces
and contexts, making it 'look' like a process which
runs inside the guest ...

btw, personally, I don't see why there should be an
additional way to logon to a guest, compared to a normal
host system, but definitely there should be a way to
do some emergency repair and testing until the guest
is perfect (at which point the enter is not required
anymore) and a quick and simple way to utilize the
guests as advanced chroot (which the enter can be
considered, although I would wish the 'running guest'
limitation would be dropped there, in util-vserver)


> Cheers
> Ed W
Received on Wed Jul 15 23:04:19 2009

[Next/Previous Months] [Main vserver Project Homepage] [Howto Subscribe/Unsubscribe] [Paul Sladen's vserver stuff]
Generated on Wed 15 Jul 2009 - 23:04:21 BST by hypermail 2.1.8