donn@utah-cs.UUCP (Donn Seeley) (01/07/87)
First, let me say where we stand. The University has purchased 6 M2361
(Super Eagle) disk drives. Of these 6, 5 have proved to be defective
in some way. One drive arrived DOA and was promptly replaced; one
drive that was purchased from a different reseller developed zillions
of hard errors on head 14 and after a few months delay it was replaced;
and we're stuck with the other three, all of which have developed hard
errors on head 1 and are past their 90-day warranty. The errors look
like scratches in that the sector number varies more or less linearly
with the track number.
We've been in touch with the various resellers and Fujitsu's American
subsidiary and we've heard some conflicting stories. A representative
of one reseller at first conceded that HDA revs D1 and D2 were bad.
Further discussions with this reseller yielded the claim that there
were no known bugs with the drive, but that some drives needed a levels
adjustment(?!). In any case we were SOL because our warranty had
expired. Another reseller told us that higher rev HDAs were not
available and that the only way they could meet our warranty was to
ship us a drive that was older than the one that broke. A
representative of Fujitsu America claimed that they had heard
indirectly of problems with the drive in the field, but that no
concrete reports had been delivered to them and that there were no
known bugs in any rev of the HDA. Basically, confusion reigns among
the resellers and Fujitsu appears to be stonewalling.
(BTW, these are my opinions and should not be construed as any sort of
official position by the University!)
Here's an edited collection of the responses:
========================================================================
Date: Wed, 3 Dec 86 12:11:40 EST
From: seismo!rick (Rick Adams)
We have 6 and they have been rock solid. No problems of any kind. We have
2 on our Celerity and 4 on various Sun workstations. All use the XY451
controller. ...
You weren't very specific about what you wanted to know, but now that
I think about it. We got 3/6 DOA. The replacements ran fine and we
have had zero problems once they were installed and running.
(The ones we got from SI were the problem drives)
---rick
------------------------------------------------------------------------
Date: Wed, 3 Dec 86 12:00:25 CST
From: seismo!harvard!husc6!cfa!wyatt (Bill Wyatt)
... Our High Energy Astrophysics group has a 2351 and a 2361.
The former has been rock-solid. The latter gave increasing errors and
lots (1-2 per week) system crashes from a month or so after installation
until about 3 weeks ago, when it had a head crash, forcing an HDA replacement.
Reformatting (before the crash) would seem to fix things for a couple of
weeks, but errors would (bad blocks) would start showing up, and the system
would crash with increasing frequency. This cycle occured twice. BTW, the
2351 is a Sun file-server. The 2361 is on a VMS 4.1, now 4.4 uVax II.
We had heard that the 2361's had problems before we installed it, but
Emulex assured us there had been a really bad batch from Fujitsu, but
ours was from after that. Emulex acknowledges there have been problems.
Our system is on a DEC service contract, and they are apparently gathering
staistics on the failure rate.
The new HDA has had no problems, but it's only been two weeks so far.
Bill UUCP: {seismo|ihnp4}!harvard!talcott!cfa!wyatt
Wyatt ARPA: wyatt@cfa.harvard.edu
------------------------------------------------------------------------
Date: Wed, 3 Dec 86 23:14:15 EST
From: jrs@ee.ecn.purdue.edu (Jeff Schwab)
We are starting to get them in here. So far no problems
what so ever. I can't say we have the years of extensive
experience as we have had with the old eagles and the cdc9766s,
but so far the large eagles seem to be a win and we are
recommending them to our users for all future purchases.
Jeff
------------------------------------------------------------------------
Date: Thu, 4 Dec 86 07:41:31 pst
From: Mark Ellson <seismo!caip!lll-lcc!leadsv!mfe>
About 3 months ago when we were starting to consider buying some M2361
drives, the distributor we work through told us they had been having
problems with the drive. I think the problem was something about the
bearing in the spindle giving out or leaking lubricant. At that time, the
distributor went back and talked to Fujitsu, and they admitted that they had
a problem but it was fixed at about that time and new drives should be fine.
From the description we had at the time, it sounded like the only possible
ECO would be to replace the HDA with a later Rev.
I hopes this helps a little.
Mark Ellson
Lockheed Missiles & Space Co.
------------------------------------------------------------------------
Date: Thu, 4 Dec 86 10:53:39 PST
From: bill@stride.Stride.COM (Bill Rainey)
Donn,
You may want to check with Ron Sheen at UNR
(stride!unrvax!ron) about the problems they had with their super
eagle. He was telling me that almost every super eagle from the first
two months of production has failed (they bought one of those) and that
they had mods to the sheet metal to improve cooling around some of the
electronics for one of the head assemblies. Their drive kept losing one
of the heads as I remember. I'm sure Ron could tell you more.
Bill
------------------------------------------------------------------------
From: genie.UUCP!scooter@cgl.ucsf.edu (Scooter Morris)
Donn,
We've got five 2361's and have had little trouble. We have recently
noticed, however that we are getting a lot of bad blocks near the
last cylinder. We also have a couple of blocks that we have been
unable to mark bad. At any rate, we're going to reformat 2 of the
drives this weekend and see what happens.
By the way, we were told by Emulex that Fujitsu did have a bad
lot of 2361's get through QC. Apparently there was some problem
which caused them to go bad very rapidly. ...
Scooter Morris
Genentech, Inc.
scooter@cgl.ucsf.edu
..ucbvax!ucsfcgl!genie!scooter
------------------------------------------------------------------------
Date: 04 Dec 86 13:38:50 PST (Thu)
From: (Ron Sheen) <unrvax!ron>
... Bill [Rainey] is right, we had lots of problems with our
super-eagle. About two months after we received it, the drive failed and
all the electronics had to be replaced. The drive work for about another
month until it failed again, this time the HDA had to be replaced. Again,
after two months it failed and the hda was replaced for the second time.
Its been running ever since (about 3 1/2 months) with no problems.
The last service engineer told me that there is some magic combination of
electronics rev and hda rev. I think the electronics rev is B2 and the
HDA rev is D, but its been a while. Apparently the HDA problems were the
result of poor aerodynmics on the bottom platters, resulting in single head
crashes that damaged the disk platter. I was told that the last rev on the
HDA is to bake the platters longer so they are harder.
It appears that grounding is a real problem at several installations. The
manual has a section on 3 - 4 different grounding techniques and it appears
that you have to trial and error the grounding until one works. Ours appears
to have worked the first time ( we have frame and signal grounds jumpered on
the drive and a #8 conductor connected to the controller ground).
I'm generally pleased with the performance and the space, we bought the drive
and controller for what we sold an RA-60 (drive only) for so I'm very happy
with the price. If the drive really is solid now, its probably going to
work out well for us. ...
-- Ron Sheen ...!seismo!unrvax!ron
------------------------------------------------------------------------
Date: Sun, 7 Dec 86 16:22:55 PST
From: ken@Hamlet.Caltech.Edu (Kenneth Adelman)
... We have one Super-Eagle and one Eagle connected via a
spectrologics 25+ controller on a MicroVAX II (VMS). We have had no
problems with the Eagle, but have lost three HDAs on the Super-Eagle
since we received it last June. The last failure was yesterday...
Ken Adelman
Caltech
------------------------------------------------------------------------
From: lacasse@randvax.UUCP (Mark LaCasse)
Date: 6 Dec 86 11:31:57 GMT
I brought up one Fujitsu M2361 "Super Eagle" here about 3 months ago.
It was on an Emulex SC7003 controller to a Vax 11/780.
The scuttlebut was that the reason our drive failed was that it
had an HDA revison number of <= D3. D4 and above is supposed to be OK.
[I could be wrong and it you could need D5 to be OK. I can't remember
if D4 is OK or not.]
The failure mode for D3 and earlier is that immediately to after six
months of perfect operation, it is prone to the seal around the main
bearing for the platters giving way. This causes bearing oil to get
in the HDA and on the platters. :-(
Our salesman said we were clear and this wasn't our problem, because
the drive said D5 on the back. (By "said" I mean the first numbers not
crossed out.) Turned out this was the rev. number of the power supply.
You have to look inside the drive to find the HDA rev. number. I think
ours was D2. Anyway, it failed.
Since Fujitsu fixed ours on warranty (HDA swap) it has been find. Too
had the HDA fix took them two months at their plant! ...
Mark LaCasse qantel!hplabs!sdcrdcf!randvax!lacasse
c/o The Rand Corporation cbosgd!ihnp4!sdcrdcf!randvax!lacasse
1700 Main Street lacasse@Rand-Unix
Santa Monica, CA 90406
213/393-0411 ext. 7420
------------------------------------------------------------------------
Date: Wed, 10 Dec 86 19:57:17 pst
From: seismo!lll-crg!lll-lcc!unisoft!rtech!rtech!cormac (Cormac Burke)
We've experienced several problems with Super Eagles. We're running them on
microvaxes under both VMS and Ultrix, and have had problems with both. Under
VMS they've written over their own bad track table, under Ultrix they seem to
have developed some bad spots in a strange manner. We're running the super
eagles with Emulex QD32 controllers, and they could just as well be the cause
of the problem. Please let me know what your experiences have been. Thanks much.
Cormac Burke
Relational Technology Inc.
(415)769-1400
sun!rtech!cormac
========================================================================
What this list shows is that a lot of people are having problems with
M2361s... Two people mention problems with failing spindle seals which
spread lubricant inside the HDA. One other suggestion is that poor
aerodynamics cause head crashes on lower platters -- the lower platters
certainly seem to give us the most trouble. It sure would be nice if
Fujitsu could give us the full story on the M2361... We aren't likely
to buy any more Super Eagles if we can't be assured that the problems
are fixed (especially if we're still stuck with several of the broken
drives and have to pay to get them replaced).
The M2361's 'design value' for MTBF is more than two years,
Donn Seeley University of Utah CS Dept donn@utah-cs.arpa
40 46' 6"N 111 50' 34"W (801) 581-5668 decvax!utah-cs!donn