[Nut-upsdev] suggested interface for an MGE Galaxy 6000
Arnaud Quette
aquette.dev at gmail.com
Fri Aug 3 09:02:41 UTC 2007
Hi Bruce,
sorry for the lag in answering...
2007/7/25, Bruce Allen <ballen at gravity.phys.uwm.edu>:
> Dear NUT Developers,
>
> My research institute is building a new cluster room, with an 800kVA MGE
> Galaxy 6000 UPS system. We will have a couple of thousand Linux nodes that
> we want to shutdown with NUT.
>
> I see that the Galaxy 6000 is available with several different interface
> options, and that NUT supports (at least) three of these interfaces:
>
> Comet / Galaxy (Serial) Utalk Serial Card (ref 66060)
> Pulsar / Comet / Galaxy (SNMP) SNMP card (ref 66062)
> Pulsar / Comet / Galaxy (SNMP)SNMP/Web Transverse card (ref 66074)
>
> I have a couple of questions:
>
> Could the NUT developers recommend which of these different interface
> options would be the best one for us to try first?
well, that's a hard choice since:
- UTalk is the old ASCII protocol
- the SNMP driver (snmp-ups) isn't complete, since it doesn't
implement the manager mode (the driver should subscribe itself with
the card SNMP Agent). The UPS so doesn't know there are connected /
protected clients, and so cut the power when it reaches the LowBattery
level.
This results in needing to handle an early shutdown on the client,
using upssched (handling on the NUT UPS server is sufficient though).
This has been used successfully on large systems...
- moreover, newer SNMP/Web card aren't well handled since their SNMP
implementation is really basic (only what is needed for Network
Management Systems, such as HP-OV). The main protocol for these cards
is the Web/XML MGE proprietary protocol. This spec hasn't been
disclosed, though nothing should stop to do so (I'm trying to finish
these kind of points before leaving for Mango / Powerware) and there
is no driver for the moment.
BTW, I'm sure that asking for the XML protocol with your setup would
help in pushing the release ;-) Contact me back privately and I'll
forward your mail.
Obviously, the driver writing is still to be done then.
So your best bet is between 66060 and 66062.
The last and important point is that I will leave the newly formed APC
| MGE in about a month for MGE OPS (the MGE small systems sold to
Powerware). And thus I won't be able to help you then :-( I can't tell
whether there will be somebody to handle this or not at APC | MGE...
> Do you anticipate any problems in scaling to ~ 2000 machines to shut down?
>
> The full load battery run time should be about 6 minutes. We will
> probably configure most machines to start a shut-down cycle if the UPS is
> online for more than about 1.5 minutes.
handling this on the NUT server side, using upssched as told above,
will be sufficient to start the clients shutdown.
The other limitation, as Arjen told, is the FD_SETSIZE.
Check this for an example on solving it:
http://www.v7n.com/forums/dedicated-servers/3129-linux-compile-apache-higher-fd_setsize.html
Arnaud
--
Free Software Developer - http://arnaud.quette.free.fr/
Debian Developer - http://people.debian.org/~aquette/
Network UPS Tools (NUT) Project Leader - http://www.networkupstools.org/
Ubuntu Media Center (UMC) Project Leader - https://launchpad.net/~umc-team
More information about the Nut-upsdev
mailing list