[Nut-upsdev] External Devices

Arjen de Korte nut+devel at de-korte.org
Sat Dec 2 21:09:28 CET 2006

Arjen de Korte wrote:


> As far as I
> can tell from your description of the patches, you'll end up with the
> same mess again if you happen to have two disk arrays serving one
> client.

Never mind the remark above, you've fixed this by naming the devices, so
you can indeed monitor multiple external devices individually.

> I would prefer a more generic method which uses NOTIFYCMD to let
> users know that the disk array is going down.

This remark still holds true.

Essentially you're duplicating the MONITOR and MINSUPPLIES from the
upsmon.conf file from the disk array to the workstations using it. I
would hate to administrate such a configuration, where you need to keep
all this in sync. Imagine adding UPSes, changes in the number of power
supplies needed (because of changed power demands) and so on.

Another thing is that I don't like the idea of powering down a client
because a service it uses is unavailable. There may be other reasons
(other than power outage on a disk array) that may mean the service is
unavailable, but can't be handled by NUT (cabling/router problems).
Powering down a client should be your last resort.

I think you're better off with a script that is called by NOTIFYCMD that
will forcibly logoff your users once power has become critical.

Kind regards, Arjen

More information about the Nut-upsdev mailing list