SUMMARY: IPC failure rate

From: Len Evens (len@math.nwu.edu)
Date: Tue Jan 28 1992 - 02:25:07 CST


Since we had experienced 2 failures for 5 IPC's purchased within the
last year, I asked what experience others had. So far a rough count
shows 55 failures for 426 machines. 30 of these were frame buffers
but one report of 25/100 biases this. 11 of the complaints are
about color monitors. Since responses were presumably biased towards
those with problems, the failure rate during the first year is
probably somewhat less than the 13% reported here. Even if it
is only 5%, is that good? What kind of failure rates should one
expect in general for a moderate sized network of different
machines?

Here are the responses so far.

==================================================================
>From sjh@helicon.math.purdue.edu Sat Jan 25 17:38:38 1992

IPCs around here have developed a reputation for trouble. Most of
ours have settled down now, but PUCC (the computing Center) has about
100 of them and had to repare one a week for a long time.
Framebuffers and disks were by far the most frequent problems.

Steve.

====================================================================
Steve Holmes \ Bitnet: sjh%math.purdue.edu@purccvm
Systems Administrator \ Internet: sjh@math.purdue.edu
Dept. of Mathematics \ uucp: ...!purdue!sjh
Purdue University, \ phone: (317) 494-6055
W. Lafayette, Indiana 47907 \ Fax: (317) 494-0548
====================================================================

>From shigeya@foretune.co.jp Sat Jan 25 18:34:54 1992

We're using two IPCs (monochrome display) for about one year and four
months. I changed configuration many time (disks, such.) but had no problem.

shigeya

>From fetrow@orac.biostat.washington.edu Sat Jan 25 20:59:48 1992

 No problems so far (2 IPC, 9 SLC).

-- 
 -dave					   fetrow@biostat.washington.edu

>From exuksm@exu.ericsson.se Sat Jan 25 21:02:09 1992

In article <9201251808.AA19004@poincare.math.nwu.edu> you write: > We have had two IPC's which were delivered last June fail >for different reasons (a framebuffer and a disk). Has anyone else >noticed any special problems with IPC's? > I don't know about the disk, but if that was a cgthree color frame buffer, weve had about 25 (out of 100) of these go bad. Apparently Sun used some "out of spec" parts when having these made. I can get the exact details if you wish. (sorry i'm at home right now)

- Karl

>From seeger@thedon.cis.ufl.edu Sun Jan 26 02:48:31 1992 Return-Path: <seeger@thedon.cis.ufl.edu>

The only thing that we have seen is a inordinate proclivity in a couple of them to corrupt their /var partitions. One did it a few times last summer, but has been good recently. The other was ok for about 3 months, 8-( but ever since will corrupt its /var whenever it is heavily loaded. We have seen this happen on a particular 3/80 (with Quantum 105) a couple of times, which made me start thinking it might be an OS bug. If it were on hardware maintenance, we would have already swapped the CPU board. As it is, we are awaiting arrival of our spare IPC CPU board.

The corruption is such that fsck will not fully fix the file system, necessitating a newfs of the partition.

The IPC with the worst problem has shown no difference in behavior after removing three external Fujitsu M2263SA drives, changing SCSI cabling, trying three different 200MB internal drives, and trying different termination combination on the now unused external SCSI port. This is not an early IPC with the SCSI bus layout problem on the CPU board. Six or seven other IPCs have not shown any problems, though.

On the other hand, about 20% of our 30 SLCs have had hardware problems, including bad Ethernet ports, bad serial/mouse ports, bad video, and defective memory expansion sockets. To date, none of our 4/65s (4) or 4/75s (9) have had any problems that come to mind.

Regards, Chuck -- Charles Seeger E301 CSE Building Office: +1 904 392 1508 CIS Department University of Florida Fax: +1 904 392 1220 seeger@ufl.edu Gainesville, FL 32611-2024 UUCP: uflorida!seeger

>From @trotter.usma.edu:aj4640@euler Sun Jan 26 07:06:42 1992

Funny you should mention it. We've two IPCs. One is fine, but the other broke 24 hours before the warranty expired! The disk went bad.

I wonder if there are enough data point do draw a decent regression line?

Jack -- _______________________________________________________________________________ Major J.S. Robertson aj4640@usma2.usma.edu Department of Mathematical Sciences (914) 938-2453 USMA, West Point, NY 10996-1786 FAX (914) 938-2409

>From svpillay@Princeton.EDU Sun Jan 26 12:22:15 1992

I had 9 IPCs delivered about two months ago. The color framebuffer failed on one within two days. Apart from that, everything has been fine.

Kanthan Pillay Systems Administrator Program in Applied and Computational Math Princeton University

Work: (609) 258-6488 Internet: svpillay@Princeton.EDU Home: (609) 396-9004 Bitnet: SVPILLAY@PUCC Fax: (609) 258-1735 uucp: princeton!svpillay

>From mikem@juliet.ll.mit.edu Mon Jan 27 08:24:10 1992

Two of the 15" color monitors (out of three) shipped with color IPCs died within a week of putting them into service. Sun replaced them under warranty and we've had no further problems with them.

------------------------------------------------------------------------------ Michael Maciolek mikem@juliet.ll.mit.edu (617) 981-3174 Network Engineer --- MIT Lincoln Laboratory Group 43 --- Weather Sensing ------------------------------------------------------------------------------ Motto: It doesn't matter how much it costs, as long as it looks cheap. ------------------------------------------------------------------------------

>From phil@pex.eecs.nwu.edu Mon Jan 27 08:43:43 1992

We have received quite a few (7 or more) and have only had one fail (bad framebuffer).

Bill

>From feldt@phyast.nhn.uoknor.edu Mon Jan 27 08:59:44 1992

Len,

No problem here, but we only have one IPC (mine!) and have only had it a little over 7 months.

Andy Feldt System Support Programmer +-------------------------------------------------+ | Andy Feldt | feldt@phyast.nhn.uoknor.edu +----------------+ Dept. Physics & Astronomy \ feldt@uokniels (BITnet) | The University of Oklahoma |UOKHEP::FELDT (HEPnet) This | * | phone: (405)-325-3961 space | 440 W. Brooks | for \ Norman, OK 73019 | rent ~\_____________________________ | ... ~'

>From chuck@fsg.com Mon Jan 27 09:38:27 1992

I installed (at a client) 3 IPC's in June - one had a bad motherboard (bad solder - a salt like substance was visible around the chips) and I forget what went wrong with the other. It happened within a week though. Unsure if there was a bad batch.

(two dozen other ones I've installed have had no problems).

Chuck -------------------------------------------------------------------- Chuck Yerkes, System Administrator Internet: chuck@fsg.com 225 Broadway 24th Floor Voice: (212) 285-8001 Fusion Systems Group Fax: (212) 285-8705 New York, NY 10007

>From miker@sbcoc.com Mon Jan 27 10:20:10 1992

We got hundreds of'em (well, nearly) and haven't seen any such problem.

>From stpeters@dawn.crd.ge.com Mon Jan 27 10:28:45 1992 Not here. We got 8 of them about last April; no failures yet.

-- Dick St.Peters, GE Corporate R&D, Schenectady, NY stpeters@crd.ge.com uunet!crd.ge.com!stpeters

>From era@niwot.scd.ucar.EDU Mon Jan 27 10:50:47 1992

We had a miserable time with a shipment of 30+ IPCs we received several months ago. Seven of the 16" monitors (the cheapie ones, with German instructions, I believe they are Dutch-made) were either totally DOA, or exhibited horizontal noise/nonlinearity/etc. We had to send these seven back. ---------- Ed Arnold * NCAR * POB 3000, Boulder, CO 80307-3000 * 303-497-1253(voice) 303-497-1137(fax) * era@ncar.ucar.edu [128.117.64.4] * era@ncario.BITNET

>From VINCE@UCONNVM.BITNET Mon Jan 27 10:55:30 1992

On Sat, 25 Jan 92 12:08:14 CST you said: We have had two IPC's which were delivered last June fail >for different reasons (a framebuffer and a disk). Has anyone else >noticed any special problems with IPC's? Yes, ours delivered last june failed with a framebuffer problem in Dec. Please summarize your results. Thanks, Vince<vince@math.uconn.edu>

>From ken@cs.sfu.ca Mon Jan 27 12:12:01 1992 Yes. We have about 10 IPC's and one monitor blew a capacitor ( 2 months after warranty). One with a disk failure during warranty and another with a framebuffer failure also during warranty. The framebuffer problem was the result of the CPU board ( caused two to fail actually ). We also have had an IPX with a trashed disk at the 3 months ,5 day mark . No sympathy from SUN. I am interested in a summary. I have a feeling that there are a lot of lemons out there.

Thank You, Ken Huckell Network Support Group Centre for System Science ken@cs.sfu.ca

>From mcgrew@cs.rutgers.edu Mon Jan 27 12:20:03 1992 Hi,

We've gotten about 10 IPC's so far, and haven't noticed any problems with 'em...

Charles

>From kwz@zabriskie.berkeley.edu Mon Jan 27 17:19:07 1992

Yes, we had a batch of 5 IPCs delivered last July of which two were very troublesome: one had a disk (Maxtor 200M internal: no response to SCSI probe) and a monitor (Sony: intermittent internal arcing sound) replaced under warranty, the other had a color frame buffer (cgthree0: complete loss of display) and the monitor (Sony: intermittent flickering brightness) replaced, also under the 90-day warranty. Since then the second one had an aftermarket SIMM fail in it, most likely a coincidence. It is unclear how these failures might be related, other than possibly mechanical stress/shock before delivery to our site. There are a total of 8 IPCs in our workgroups.

kevin

-------------------------------------------------------------------- Kevin Zimmerman 384 Cory Hall System Support Berkeley CA 94720 U.C. Berkeley (510) 642-0152 Electronics Research Laboratory kwz@zabriskie.berkeley.edu --------------------------------------------------------------------



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:35 CDT