[Beowulf] KVM to a compute node - ssh

Robert G. Brown rgb at phy.duke.edu
Thu Jun 3 09:19:40 PDT 2004


On Thu, 3 Jun 2004, Guy Coates wrote:

> >
> > PCs, even "server class" motherboard PCs, alas, have never been designed
> > to have a "real" serial console as video cards are viewed to be
> > essential, not optional.
> 
> > image.  This is a BETTER solution than a serial interface for this
> > purpose, as serial interfaces to my experience have a real problem
> > interpreting e.g. three finger salutes in all possible runtime states
> 
> Ok, I admit plain serial does have its limitations on PC hardware.

Ya, and WOL isn't perfect or even all that great for this as well -- it
isn't really designed to overcome post boot state from what I can tell
from the original AMD white paper, and so while it can power a system on
from off, it isn't clear that it can reboot a hung system.  In a way,
one would fear this as it would enable a truly evil DoS attack if one
could force a system reboot by sending it an arbitrary packet.  OTOH, it
would be a powerful force for Good in the hands of the righteous.  Like
so many things.

> >made something of an end run around the problem, what I'd REALLY like at
> > this point is a true network-based console -- one where one can connect
> > to a brand new box's bios and console via the network.
> 
> That holy grail does just about exist.  I was plesantly surprised with the
> latest intel server offering from HP.  As well as a serial port it comes
> with an integrated management processor (ILO), basically just another rj45
> port which you plug into your LAN. You can talk to the ILO via telnet or a
> web interface, and do remote power, hard reset etc by firing carefully
> crafted bits of XML at it.

Yes, this would do very nicely, although I don't really see why it needs
to consume a SECOND network interface and switch port.

> 
> The management processor is pretty clever; if it notices something on the
> ILO or serial port it automatically changes the BIOS interface into its
> command line version, and spits all its post stuff etc onto the ILO or
> serial port. You can set user defined escape characters to send those
> awkward CTRL-ALT-DELETE or magic SYSRQ sequences.
> 
> You get the best of both worlds; command line driven, out of band
> management, a real console, but going across the network rather than
> through an expensive terminal server.
> 
> Drawbacks; it is HP only, and I don't know if it is IPMI compliant. A
> similar management interface that is the same on *all* pc hardware would
> be even better, but it is at least a start.

Absolutely.  Even with our server/cluster room in the building, it is a
pain to walk down there every time something needs to be bounced (like
right now:-) and sometimes I work at home (like right now).  WOL won't
bounce them, at least as far as my experiment went today.  I may have to
meddle in the BIOS first.

   rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu






More information about the Beowulf mailing list