Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!PUMPKIN-PATCH.NOSC.MIL!dennis
From: dennis@PUMPKIN-PATCH.NOSC.MIL (Dennis Cottel)
Newsgroups: comp.sys.apollo
Subject: Re:  MTBFs of Apollos
Message-ID: <8904281433.AA26532@cod.nosc.mil>
Date: 28 Apr 89 14:32:31 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 44

Rich Blinne asked:

> My manager is working on a maintenence study for our Apollos and
> needs to know the mean time between failures for various different
> types of Apollos.  Does anyone know the MTBFs for Apollo 3000, 3500,
> 4000, and 4500s.  Thank you.

Last year in response to a similar internal request, I put together
the following note.  My feeling is that things have been at least as
good -- maybe better -- since then.
                                                          --Dennis

	Dennis Cottel  Naval Ocean Systems Center, San Diego, CA  92152
	(619) 553-1645      dennis@nosc.MIL      sdcsvax!noscvax!dennis



--------------------------------------------------------------------------
Hardware reliability summary for DN3000/DN4000 machines.     July 13, 1988
--------------------------------------------------------------------------
We have 20 DN3000s and 1 DN4000.

We now have 266 node-months of use out of these machines.

I ignored hardware calls that were made during the first few weeks
which seemed to have to do with getting the machine working right in
the first place.

We have had 13 hardware failures:
   4 displays replaced
   1 display controller card replaced
   2 noisy disks - fixed without replacing disks    *
   2 tape controller cards loose                    * 
   1 power supply replaced
   1 tape drive replaced
   1 disk controller card replaced
   1 new battery

The MTBF is therefore about 20 node-months.

If you don't count the *starred items above which were not really
hardware failures in that nothing had to be replaced, then there
are 9 broken pieces over the 266 node-months -- pretty close to
three years MTBF.