Is there a "alternative" way of doing kvm-over-IP
Buying the hardware boxes sounds pretty attractive, but I wondered if any industrious open sourcers have come up with anything that any NANOG-ers have come across that would do something similar to what the Dell 2161DS box would do. I realize we would need a server with some cards in it to terminate the connections and then some cards to go in the boxes, we just need a bit more density than 16 ports. Tia -Drew
On Wed, 20 Jul 2005, Drew Weaver wrote:
Buying the hardware boxes sounds pretty attractive, but I wondered if any industrious open sourcers have come up with anything that any NANOG-ers have come across that would do something similar to what the Dell 2161DS box would do. I realize we would need a server with some cards in it to terminate the connections and then some cards to go in the boxes, we just need a bit more density than 16 ports. There is an open-sores project to design KVM-over-IP, at http://okvm.sourceforge.net/kvmoverip.html
However, all you want is Dell 2161DS PEM (port extender module, equivalent of avocent ARI). With PEM, you can attach 8 devices to each of the 16 ports of the 2161DS, total 128 ports per 2161DS. -alex
On Wed, 20 Jul 2005, Drew Weaver wrote:
Buying the hardware boxes sounds pretty attractive, but I wondered if any industrious open sourcers have come up with anything that any NANOG-ers have come across that would do something similar to what the Dell 2161DS box would do. I realize we would need a server with some cards in it to terminate the connections and then some cards to go in the boxes, we just need a bit more density than 16 ports.
I've been mostly successful in eliminating the need for most of our kvm ports, formerly with pc weasels, and more recently with servers supporting ipmi directly or through an addon card. the pc weasels required terminal servers the ipmi cards just require that you put the first network interface on the machine someplace you can reach. weasel http://www.realweasel.com/ ipmi http://www.intel.com/design/servers/ipmi/
Tia
-Drew
-- -------------------------------------------------------------------------- Joel Jaeggli Unix Consulting joelja@darkwing.uoregon.edu GPG Key Fingerprint: 5C6E 0104 BAF0 40B0 5BD3 C38B F000 35AB B67F 56B2
On Wed, Jul 20, 2005 at 04:53:36PM -0400, Drew Weaver wrote:
Buying the hardware boxes sounds pretty attractive, but I wondered if any industrious open sourcers have come up with anything that any NANOG-ers have come across that would do something similar to what the Dell 2161DS box would do. I realize we would need a server with some cards in it to terminate the connections and then some cards to go in the boxes, we just need a bit more density than 16 ports.
If you just need KVM to a number of Unix systems, you can use remote X-Windows logins via 'xdm'. Or just 'ssh'. This does not serve the problem of using the remote console when the system is not up and running the X client programs. If you want console access whether up or down, and are willing to settle for serial console access, Cyclades has stackable boxes with 48 ports each, and there exist cards ("PC Weasel", <http://www.realweasel.com/>) that allow you to work with PC systems whether they are up or down, as if on the video console. The Dell 2161DS can stack to up to 16 X 16 ports. I realize this doesn't change the density, but it might help. -- Joe Yao ----------------------------------------------------------------------- This message is not an official statement of OSIS Center policies.
On Wed, 20 Jul 2005, Joseph S D Yao wrote:
If you want console access whether up or down, and are willing to settle for serial console access, Cyclades has stackable boxes with 48 ports each, and there exist cards ("PC Weasel", <http://www.realweasel.com/>) that allow you to work with PC systems whether they are up or down, as if on the video console.
I'd second the notion on the PC Weasel. I know the guy who designed them (hpeyerl), and they were designed from the start to be indistinguishable to the OS from textmode VGA cards and PS2 keyboards. The redraw algorithm is smart, along the lines of "screen" -- some serial BIOS support I've seen is far too full-screen-redraw happy (<cough>Dell<choke>). -- -- Todd Vierling <tv@duh.org> <tv@pobox.com> <todd@vierling.name>
I'd second the notion on the PC Weasel. I know the guy who designed them (hpeyerl), and they were designed from the start to be indistinguishable to the OS from textmode VGA cards and PS2 keyboards. The redraw algorithm is smart, along the lines of "screen" -- some serial BIOS support I've seen is far too full-screen-redraw happy (<cough>Dell<choke>).
Thirded, Herb did a stunning job on these.
participants (6)
-
alex@pilosoft.com
-
Drew Weaver
-
Joel Jaeggli
-
Joseph S D Yao
-
Neil J. McRae
-
Todd Vierling