Subject: Re: RAID-5 benchmark results
To: Gavan Fantom <>
From: Robert Elz <kre@munnari.OZ.AU>
List: netbsd-users
Date: 12/14/2001 22:25:01
    Date:        Fri, 14 Dec 2001 12:26:39 +0000 (GMT)
    From:        Gavan Fantom <>
    Message-ID:  <>

  | I'm sure I'm not the only person on this list who has seen all of the
  | above happen.

Melbourne University simply disconnected almost all of the UPSs that
were in use, because they failed more frequently than the line power
did, and caused more disruption when they failed (external power will
generally simply come back after a few seconds/minutes/hours/..., a dead
UPS tends to need actual attention to get the system(s) connected back
running again).

That is, when a UPS is in use, it becomes the critical resource - it
isn't in the least redundant, UPS fails and system dies, regardless of
the current state of the external power grid (we have a few devices with
dual redundant power supplies, with one connected to UPS, the other to
line power, which then goes give some redundancy - but your average
scsi drive, which is where this all started, doesn't have that kind
of option).


ps: This isn't to say that the UPSs failed all that often - more that
the external power was even more reliable most of the time.