Sun-Spots-Request@RICE.EDU (William LeFebvre) (08/16/88)
SUN-SPOTS DIGEST Monday, 15 August 1988 Volume 6 : Issue 184
Today's Topics:
Re: 892MB disks shudder during fsck (2)
Re: Bug with cc -M (2)
Re: making yp do name server lookups
Re: Sun X
Re: Tape Drive Allocator program?
Re: clearing the 68020 on-chip instruction cache
German Sun User Group founded
Patch for SUN OS 4.0 ALM-2 bugs
Sun ALM2 Problem (Ciprico Prob?)
ALM2 problem confirmed
strange timing problem in /usr/ucb/Mail
SCSI disk on a SUN-2?
Mail readable from all yp clients?
Send contributions to: sun-spots@rice.edu
Send subscription add/delete requests to: sun-spots-request@rice.edu
Bitnet readers can subscribe directly with the CMS command:
TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name
Recent backissues are available via anonymous FTP from "titan.rice.edu".
For volume X, issue Y, "get sun-spots/vXnY". They are also accessible
through the archive server: mail the request "send sun-spots vXnY" to
"archive-server@rice.edu" or mail the word "help" to the same address
for more information.
----------------------------------------------------------------------
Date: Tue, 9 Aug 88 22:02:40 EDT
From: scs%lokkur.UUCP@umix.cc.umich.edu (Steve Simmons)
Subject: Re: 892MB disks shudder during fsck (1)
Reference: v6n171
In response to...
>The most likely reason for the thrashing is that /etc/fstab contains 2 or
>more entries on that one drive with the same pass number:
>
> |
> V
>/dev/xy0a / 4.2 rw,noquota 1 1
>/dev/xy0b /usr 4.2 rw,noquota 1 1
While it's great to know what that last field is for, there's probably a
much more prosaic contributing factor -- either inside or out, there is
something physically loose in the drive. Perhaps the HDA is not correctly
attached to the mount, perhaps the drive isn't mounted correctly in the
cabinet, perhaps the cabinet permits too much flex. In any case SHAKING
UP AND DOWN 1/4 INCH IS NOT NORMAL!!!! It is very likely to cause
catastrophic failure (ie, head crash). I have never seen a properly
mounted drive move even 1/16 of an inch, and certianly not up and down!
Head motion is horizontal, not vertical (yes, I know horizontal forces can
be transformed. phhht! :-) )
And Larry is certianly right -- the funky fsck is causing the excessive
activity. But the excessive activity should not create that sort of
motion in a properly mounted drive.
------------------------------
Date: Tue, 9 Aug 88 14:57:41 PDT
From: chessin@sun.com (Steve Chessin)
Subject: Re: 892MB disks shudder during fsck (2)
>/dev/xy0a / 4.2 rw,noquota 1 1
>/dev/xy0b /usr 4.2 rw,noquota 1 1
>
>Both of these partitions will be fsck'd simultaneously.
Not quite true.
All filesystems with a pass number of 1 are checked sequentially. All
filesystems with a given pass number greater than one are checked
simultaneously during that pass.
[[ Gee. I guess the manual page must be wrong then. 3.5 manual page says
"All file systems with mnt_passno of 1 are first checked simultaneosly,
then all file systems with mnt_passno of 2, and so on." 4.0 manual page
says "Filesystems with the same pass number are checked simultaneously..."
No mention is made of handling pass==1 as a special case. --wnl ]]
I set up my fstab files as follows:
1. / is in pass 1 by itself (but see below).
2. For subsequent pass numbers, I choose one filesystem from each disk,
selecting them so that they are comparable in size.
This seems to give the best performance.
[[ Perhaps it is the case that fsck will always check "/" separately
during a pass. This would appear to give the behavior you described
earlier. But I cannot find any mention of that in the manual pages (and
we don't have source). --wnl ]]
Note: in 4.0, /usr is mounted read-only by /etc/rc.boot before the fsck,
and thus should also be given a pass number of 1. Both it and / are
remounted read-write if the fsck is successful. (/ was mounted read-only
by the kernel unless the boot flag -w was used.)
--Steve Chessin
Sun Microsystems
chessin@sun.com or ...!sun!chessin
(These are my opinions only, and not necessarily those of Sun.)
------------------------------
Date: Mon, 8 Aug 88 15:48:09 EDT
From: Root Boy Jim <rbj@nav.icst.nbs.gov>
Subject: Re: Bug with cc -M (1)
> Description:
>...
> The problem is that source file names are printed on stdout
> when the -M (make dependency) flag is used with more than
> one source file is specified. This is incorrect.
>...
I noticed this too, but I don't understand the objection. Unless you
generate your .c files from another suffix by an implicit rule (or use the
.DEFAULT entry to say, co them), they seem harmless.
Since mkdep is a shell script, you can grep out the offending lines.
(Root Boy) Jim Cottrell <rbj@icst-cmr.arpa>
National Bureau of Standards
Flamer's Hotline: (301) 975-5688
The opinions expressed are solely my own
and do not reflect NBS policy or agreement
------------------------------
Date: Mon, 08 Aug 88 23:28:46 PDT
From: Craig Leres <leres@helios.ee.lbl.gov>
Subject: Re: Bug with cc -M (2)
Root Boy Jim writes:
>I noticed this too, but I don't understand the objection. Unless you
>generate your .c files from another suffix by an implicit rule (or
>use the .DEFAULT entry to say, co them), they seem harmless.
Both of the reasons you suggest are sufficient to object to this bug.
An even better reason is that the the program's behavior is contrary to
the documentation:
-M Run only the macro preprocessor on the named C
programs, requesting that it generate makefile
dependencies and send the result to the standard
output (see make(1) for details about makefiles
and dependencies).
I've read that paragraph several times and I haven't found anything about
gratuitously spewing source file names.
>Since mkdep is a shell script, you can grep out the offending lines.
This is an unnecessary kludge since it's possible to make /lib/compile
work correctly.
Guy Harris of Sun tells me that he also found the bug and that it's fixed
in 4.0.
Craig
P.S. Pardon the flame, but 32 lines of included text, 6 lines of
signature, and only 5 lines of "new" text? If I thought my article was
THAT important, I would have posted it twice.
[[ That's one reason this list is moderated: I routinely trim down
text that is quoted from previous articles. --wnl ]]
------------------------------
Date: 9 Aug 88 15:29:37 GMT
From: Sven-Ove Westberg <sow@cad.luth.se>
Subject: Re: making yp do name server lookups
| "makedbm -b" adds a magic entry to your hosts database, YP_INTERDOMAIN or
| some such, that tells ypserv that if it can't find a host in the database,
| go off and try the nameserver....
On a real Unix host designed for Internet usage (eg 4.3bsd) I can do this
in a much better way. Did Sun support this also??
HOSTNAME(7) UNIX Programmer's Manual HOSTNAME(7)
NAME
hostname - host name resolution description
------------------------------
Date: 9 Aug 88 14:56:10 GMT
From: Alan Henry <mcvax!torch.co.uk!alan@uunet.uu.net>
Subject: Re: Sun X
As regards X on Sun workstations, Torch are currently selling binary
versions of X for the Sun 3/x60 series of workstations with, I believe the
latest patches and some internal bug fixes applied especially to the
colour server. For prices and details contact:
Martin Harris
Torch Computers
Abberley House
Great Shelford
Cambridge CB2 5LQ
UK
'phone (UK) 0223 841000
- Alan
------------------------------
Date: 9 Aug 88 18:04:13 GMT
From: cudcv%WARWICK.AC.UK@cunyvm.cuny.edu (Rob McMahon)
Subject: Re: Tape Drive Allocator program?
In article <DIGEST Volume 6 : Issue 171> bovet@hao.ucar.edu (Ray Bovet) writes:
>I am looking for code to allocate and deallocate tape drives.
I have a program called `with' that does most of this, and which has
served us well for a long time. It was sent to comp.sources.unix some
time ago, have a look back through the archives. It allocates a drive
under a separate shell, so that it is automatically released when the
shell exits. It will optionally send a message to the operators to load a
particular tape. It will tell you who has the drive. It will sort of
queue requests (not in the background though, it'll just sit and wait
until it's available, but you can do things like
with tape "Rob's No.1 tape, please" -c "tar xvp; mt offl" >& tar.log &
). The only thing mentioned it doesn't do is remind the person he has a
drive. Maybe it could change the prompt ?
If you can't get to the archives, I'll try to send you a copy, but I've
had bad experiences trying to send C programs to the States (all the $'s
get turned into G and $ and things ...)
Rob
--
UUCP: ...!mcvax!ukc!warwick!cudcv PHONE: +44 203 523037
JANET: cudcv@uk.ac.warwick ARPA: cudcv@warwick.ac.uk
Rob McMahon, Computing Services, Warwick University, Coventry CV4 7AL, England
------------------------------
Date: Tue, 9 Aug 88 15:28:01 EDT
From: haahr%bogey.Princeton.EDU@princeton.edu
Subject: Re: clearing the 68020 on-chip instruction cache
I asked in sun-spots volume 6, issue 170 if there was a way to clear the
68020's instruction cache from user level on the sun-3. This is necessary
for using self-modifying code, and, it turns out, for Sun's implementation
of shared libraries.
Doug Ward (daw@sun.com) and Matt Jacob (mjacob@sun.com) both pointed me at
trap #2, which does exactly what I needed. Neither thought that it was
documented, though daw noted that the code is in locore.s.
paul haahr
princeton!haahr or haahr@princeton.edu
------------------------------
Date: Tue, 9 Aug 88 19:53:23 +0200
From: unido!DFVLRGO1.BITNET!TS36@uunet.uu.net
Subject: German Sun User Group founded
A Sun User Group has been founded in West-Germany.
Activities are: 1.) distribution of newsletters
2.) e-mail list: sug-info-request@pbinfo.uucp
3.) distribution of public domain software
4.) first conference on Sept.26th 1988
to be held in Hannover (West-Germany)
5.) a lot more
People (on sites in germany) who are interested may subscribe to the list
to learn more. Feel free to ask me for further information.
------------------------------
Date: Wed, 10 Aug 88 15:03:08 EDT
From: weltyc@fs3.cs.rpi.edu (Christopher A. Welty)
Subject: Patch for SUN OS 4.0 ALM-2 bugs
SUN has informed me that there is a serious bug in 4.0 device drivers for
ALM-2's that cause modem lines to not work and printer ports to crash the
system. I have made the patch available via anonymous ftp at
turing.cs.rpi.edu (128.213.5.90) in pub/patch.tar.Z. The file is big (2
Meg compressed). The tape is also available from SUN, and was `supposed'
to be shipped with 4.0 to people with ALM-2 configurations (oh yeah,
right). I have applied this patch and it works (after one small fix which
is in the stuff I made available).
Christopher Welty --- Asst. Director, RPI CS Labs
weltyc@cs.rpi.edu ...!rutgers!nysernic!weltyc
[[ Note that this patch probably does not fix the problem described
below: --wnl ]]
------------------------------
Date: 9 Aug 88 14:58:15 GMT
From: emory!km@gatech.edu (Ken Mandelberg)
Subject: Sun ALM2 Problem (Ciprico Prob?)
We are running SunOS 4.0 on a Sun 4/280. We are having a serious problem
with our ALM2. After a few hours from boot the ALM2 goes into what I call
"history" mode. That means that when a new character is received by the
ALM2 it sends an old character to the CPU (maybe with a 20 character lag).
Each keystroke forces out one of the old waiting characters. From time to
time it is also possible for the whole machine to hang until a new
character is sent to the ALM2.
Sun has some ALM2 patches for SunOS 4.0, but none of them fix our problem,
nor does our problem seem to match the abstracts of the known problems.
My interpretation of the symptoms is that some of the ALM2 interrupts are
being lost.
The device worked properly under SunOS 3.2, so we don't think its
hardware. Since we went to SunOS 4.0, we also have added a Ciprico 3200
disk controller with NEC drives, to supplement an existing Xylogics 451
with Fujitsu drives. The Ciprico is the only non-Sun hardware and of
course is a bit suspect.
Has anyone had a similar problem, or have any ideas? I would like to hear
from anyone else running a Sun 4 under SunOS 4.0 with a Ciprico and ALM2.
Ken Mandelberg | {decvax,sun!sunatl,gatech}!emory!km UUCP
Emory University | km@emory BITNET
Dept of Math and CS | km@emory.ARPA ARPA,CSNET
Atlanta, GA 30322 | Phone: (404) 727-7963
------------------------------
Date: 9 Aug 88 19:41:54 GMT
From: harry@rainbow.atmos.washington.edu (Harry Edmon)
Subject: ALM2 problem confirmed
In comp.periph, Ken Mandelberg writes:
> We are running SunOS 4.0 on a Sun 4/280. We are having a serious problem
> with our ALM2. After a few hours from boot the ALM2 goes into what I call
> "history" mode....
I also have a Ciprico controller (3220) and an ALM2 running SunOS 3.5 on a
Sun 3/160 with the same problem. Does anyone have a clue as to the source
of this problem?
Harry Edmon ARPA: harry@rainbow.atmos.washington.edu
(206) 543-0547 UUCP: uw-beaver!geops!elnino!harry
Department of Atmospheric Sciences BITNET: HARRY@UWARITA
University of Washington
------------------------------
Date: Tue, 9 Aug 1988 14:44:00 CDT
From: King Ables <ables@mcc.com>
Subject: strange timing problem in /usr/ucb/Mail
We noticed a strange timing problem in /usr/ucb/Mail today that we've been
seeing for some time (so I suspect it exists in MANY versions). We are
currently running SunOS 3.5 on the machine where the problem occurs, but
have observed *at least* as far back as 3.0. I suspect it exists from the
original Berkeley code.
The problem occurs when replying to a message which many recipients and
you ^D at the end of the message and the CC: line is shown to you (you
have askcc set in your .mailrc). If the Cc: line has several addresses on
it so that it wraps to a 2nd line AND if you hit ^D and <RETURN> (the
<return> in response to the Cc: line prompt) VERY QUICKLY, there seems to
be a timing problem where the 2nd line's worth of text (the part that was
wrapped) isn't gathered up and processed so you wind up with a partial CC:
line going out. This has the effect of possibly leaving some people off
the Cc: line and usually cutting some address in half so that it no longer
is a valid address (half a hostname or some such).
Has anyone noticed this before and has anyone ever done anything about it?
I don't want to go figuring it out for myself is someone already has fixed
it.
Thanks for any info.
-king
ARPA: ables@mcc.com
UUCP: {gatech,nbires,seismo,ucb-vax}!cs.utexas.edu!pp!ables
------------------------------
Date: Tue, 9 Aug 88 11:37:07 edt
From: tarsa@abyss.dec.com (Greg Tarsa)
Subject: SCSI disk on a SUN-2?
I have a Micropolis 1325 disk with integral SCSI controller. Is there any
way I can get my Multibus Sun-2 to run with it? I tried the obvious, but
cannot get DIAG to format, or even recognize the disk.
Any help?
Greg Tarsa
Tarsa Software Consulting
33 Seabee Street
Bedford, NH 03102
(603)668-8349 {decuac,decvax}!elijah!tarsa
------------------------------
Date: Sat, 9 Jul 88 17:56:20 EDT
From: brickman@cme-durer.arpa (Jonathan E. Brickman)
Subject: Mail readable from all yp clients?
Does anyone know how to allow any Sun off of one yp server access mail
destined for one particular client? We have 15-20 Suns operating where 5
or 6 are configured as servers and the rest as clients to all of the
above. Currently, all mail per user can only be read from exactly one of
the Suns, as configured on the (single) yp server. Is there any way
around this? I'd LOVE to be able to read my mail from any of our
networked Suns containing my username...
Jonathan E. Brickman
(brickman@norman0.cme.nbs.gov)
------------------------------
End of SUN-Spots Digest
***********************