From MOON at MIT-MC  Wed Oct 28 00:00:00 1981
From: MOON at MIT-MC (MOON at MIT-MC)
Date: 28 Oct 1981 00:00
Subject: Installing TX XTEX
Message-ID: <mailman.495.1479821112.4695.its@nocrew.org>

This program doesn't work with files longer than 128K words (due to
overflow problems in a halfword).  It should be rewritten, for this
and numerous other reasons.  Volunteers?

From SWG  Tue Oct 27 00:00:00 1981
From: SWG (S. W. Galley)
Date: 27 Oct 1981, 00:00
Subject: [Re: catatonia - ACs/ JRST 4,]
In-Reply-To: Message of 21 Oct 81 at 1502 EDT by MOON@MIT-MC
Message-ID: <[MIT-DMS].213675>

The PI lights at the time indicated normal states, I believe:
	PI IN PROGRESS [none]
	PI REQUEST [none]
	PI ACTIVE 1-7
	IOB PI REQUEST 2,3,7
Curiously, the parity buffer didn't match the instruction register:
	INSN+MA/  344000,,17 [AOJA 17]
	PAR BUFF/ 354000,,17 [AOSA 17], ODD, STOP
I suppose your theory about some glitch turning off the PI system
is a good one.


From RLB at MIT-MC  Tue Oct 27 00:00:00 1981
From: RLB at MIT-MC (RLB at MIT-MC)
Date: 27 Oct 1981 00:00
Subject: No subject
Message-ID: <mailman.496.1479821112.4695.its@nocrew.org>

MC's T300 lossages of 7pm this evening noted in system log
occurred while I had an Emacs writing lots of blocks to FOURTH:,
like 10000 :ew... 100<hp>


From Moon at MIT-MC  Mon Oct 26 00:00:00 1981
From: Moon at MIT-MC (Moon at MIT-MC)
Date: 26 Oct 1981 00:00
Subject: System crashing works differently
Message-ID: <mailman.497.1479821112.4695.its@nocrew.org>

ITS 1243, which has been installed on MC for a week or so now, does its
crashing differently.  Most halts (JRST 4,.) have been replaced with
the new BUG macro, which types a message and goes to DDT.  Documentation
on this macro may be found in the source, where it is defined, near the
front of ITS >.  The BUG macro also replaces the SYSMSG facility.

When this system is installed on other machines, it would be a good idea
to install a new non-timesharing DDT along with it.  Assemble MC:SYSTEM;DDT.
The only change is to enable the feature of typing <alt>P to revive the
system from a non-fatal error to work.  Also install MC:SYS:TS SYSMSG when
and where the new system is installed.

From CSTACY  Sun Oct 25 00:00:00 1981
From: CSTACY (Christopher C. Stacy)
Date: 25 Oct 1981, 00:00
Subject: BUG => its
Message-ID: <[MIT-DMS].213553>


The DOVER program on DM thinks there is a chaosnet here
and tried to use undefined system calls when it cannot
spool to MC. It should just bomb out.



From MOONatMIT-MC  Fri Oct 23 00:00:00 1981
From: MOONatMIT-MC (David A. Moon)
Date: 23 October 1981, 00:00
Subject: ITS IMP lossage
Message-ID: <mailman.498.1479821112.4695.its@nocrew.org>

    Date: 23 Oct 1981 1026-EDT
    From: J. Noel Chiappa <JNC at MIT-XX>
    Subject: ITS IMP lossage
    To: jan at MIT-XX, bug-its at MIT-AI
    cc: JNC at MIT-XX

	    When an IMP goes down and is reloaded, ITS does not
    do the right thing in terms of bringing the net back up.
    It's necessary to go do a LOCK/N by hand to bring it up.
This is not actually true.  The time when ITS does not bring it back
up is when it goes down several times in rapid succession (within
30 seconds I think), after which it decides that it must be broken
and leaves it down.

From MOON at MIT-MC  Wed Oct 21 00:00:00 1981
From: MOON at MIT-MC (MOON at MIT-MC)
Date: 21 Oct 1981 00:00
Subject: catatonia - ACs/ JRST 4,
Message-ID: <mailman.499.1479821112.4695.its@nocrew.org>

    Yesterday DM entered a catatonic state: in user mode, only the null job
    running, no response to ^Z or the left-most data switch.
    I stopped it with the STOP key and dumped it to DM:.;CRASH 6089.
    The ACs were clobbered: 0/ JRST 4,325  1-16/ JRST 4,  17=U/ AOJA U.
    XUUOH/ 700000,,STYO+1.  Truly strange!

Those are the correct ACs for the null job.  Everything in the crash dump
looks completely normal.  You didn't say what the PI lights indicated;
perhaps a power surge or static electricity turned off the PI system or
stopped the machine or otherwise made it not take interrupts.



From JNCatMIT-XX  Fri Oct 23 00:00:00 1981
From: JNCatMIT-XX (J. Noel Chiappa)
Date: 23 Oct 1981, 00:00
Subject: ITS IMP lossage
Message-ID: <mailman.500.1479821112.4695.its@nocrew.org>

	When an IMP goes down and is reloaded, ITS does not
do the right thing in terms of bringing the net back up.
It's necessary to go do a LOCK/N by hand to bring it up.
I'm too burned out to grovel over the code and figure out
what it does do and should do; perhaps someone would like
to tackle this?
-------

From SWG  Wed Oct 21 00:00:00 1981
From: SWG (S. W. Galley)
Date: 21 Oct 1981, 00:00
Subject: catatonia - ACs/ JRST 4,
Message-ID: <[MIT-DMS].213236>

Yesterday DM entered a catatonic state: in user mode, only the null job
running, no response to ^Z or the left-most data switch.
I stopped it with the STOP key and dumped it to DM:.;CRASH 6089.
The ACs were clobbered: 0/ JRST 4,325  1-16/ JRST 4,  17=U/ AOJA U.
XUUOH/ 700000,,STYO+1.  Truly strange!


From MMCMatMIT-AI  Wed Oct 14 00:00:00 1981
From: MMCMatMIT-AI (Mike McMahon)
Date: 14 October 1981, 00:00
Subject: No subject
Message-ID: <mailman.501.1479821112.4695.its@nocrew.org>

SYSTEM;FSDEFS 36 is broken.  UDBLKS is multiply defined.

From CSTACYatMIT-AI  Mon Oct 12 00:00:00 1981
From: CSTACYatMIT-AI (Christopher C. Stacy)
Date: 12 October 1981, 00:00
Subject: Jobs magically appearing (Robert P. Krajewski <UC.RPK at MIT-EECS>)
Message-ID: <mailman.502.1479821112.4695.its@nocrew.org>

I responded to this person.



From UC.RPKatMIT-EECS  Mon Oct 12 00:00:00 1981
From: UC.RPKatMIT-EECS (Robert P. Krajewski)
Date: 12 Oct 1981, 00:00
Subject: Jobs magically appearing
Message-ID: <mailman.503.1479821112.4695.its@nocrew.org>

When I came over to AI via the CHAOSnet (from MIT-EECS), I decided to list my
jobs BEFORE logging in (for no apparent reason).  This is what I got:

$$v
* CONIVR P 63
  EMACS P 24
  MAIL P 25
  LISP R 5
  FOO - 17
  MACSYM R 34
  PLANER R 45
  DIRECT P 57
  UNIVERSE-SIMULATION W 107
*:version
AI ITS.1223. PWORD.2449.
TTY 42
16. Lusers, Fair Share = 29%

*:time
The time is 15:45:36 EDT.
Today is Monday, the 12th of October, 1981.
AI ITS 1223 has run for 1 day, 15 hours, 45 minutes, 28 seconds.
:KILL
*$$u

AI ITS 1223  Console 42 Free. 15:45:46


The only thing I did before logging in was a finger.  How can this be ?
I noticed that AI PWORD's appearance had changed a little.

bob
-------


From KLOTZ at MIT-MC  Wed Oct  7 00:00:00 1981
From: KLOTZ at MIT-MC (KLOTZ at MIT-MC)
Date: 07 Oct 1981 00:00
Subject: --More--
Message-ID: <mailman.504.1479821112.4695.its@nocrew.org>

In the past couple of weeks, I've noticed that --more-- breaks have done some
unusual things.  I don't remembere then doing this before, but I'm not sure.

Just now, a proceeded job tried to print out something, while my hactrn was in
the middle of printing a page (of mail, read with 
).  When the --more--
appeared, it acted as if it got a space immediately and went to the next page.

This was just now on a Vt52 on mc.  Last week, I was using a dissapoint over a
dialup, and every CLI interrupt would cause --more-- to flush.  This may be the
disappoint being random, but it didn't happen any other time.

 seems to have a problem with --more--, too.  A couple of times I've typed 
while in the middle of a page and gotten "--more--flushed
" printed out.  This was on a dialup.

Leigh.


From RLB at MIT-MC  Wed Oct  7 00:00:00 1981
From: RLB at MIT-MC (RLB at MIT-MC)
Date: 07 Oct 1981 00:00
Subject: No subject
Message-ID: <mailman.505.1479821112.4695.its@nocrew.org>

I see I dumped the earlier MC crash to .;CORPG0 4
Handwriting in the log doesn't make it to BUG-ITS, does it?


From KLHatMIT-AI  Tue Oct  6 00:00:00 1981
From: KLHatMIT-AI (Ken Harrenstien)
Date: 6 October 1981, 00:00
Subject: > name conventions
Message-ID: <mailman.506.1479821112.4695.its@nocrew.org>

That actually sounds like a good idea.  I don't know if anyone is
still keeping a list of ITS things-to-do, though.

From GLRatMIT-AI  Tue Oct  6 00:00:00 1981
From: GLRatMIT-AI (Jerry Roylance)
Date: 6 October 1981, 00:00
Subject: > name conventions
Message-ID: <mailman.507.1479821112.4695.its@nocrew.org>


The naming conventions for an FN1 or FN2 of > could be changed
to allow file extentions and version numbers.

Say my directory has these files:

	DSK:FOO; BAR FASL
	DSK:FOO; BAR LSP123
	DSK:FOO; BAR LSP124
	DSK:FOO; BAR PRESS
	DSK:FOO; BAR TXT256
	DSK:FOO; BAR TXT257
	DSK:FOO; BAR UNFASL

Then file lookup (for read) should work as:

   BAR >       refers to BAR TXT257 (normal)
   BAR TXT>    refers to BAR TXT275
   BAR LSP>    refers to BAR LSP257


			Jerry

From CStacyatMIT-AI  Sun Oct  4 00:00:00 1981
From: CStacyatMIT-AI (Christopher C. Stacy)
Date: 4 October 1981, 00:00
Subject: Ten/11 Interface seems to be flakey (so, what else is new?)
Message-ID: <mailman.508.1479821112.4695.its@nocrew.org>

About midnight, AI stopped with a parity error. When CONTinued,it
would halt again. During this time, the door buzzer was stuck on. 

Cycling the PDP-11 got the door to stop being noisy.
However, I could not get the TVs work on ITS again.

While I was scratching my head, ITS crashed at CHACK1+14.
This is in some CORE code, and the error had something
to do with a pointer to a non-existant luser.

About now, I decided things were probably screwed up enough to
want to reload and try again.  Reloaded. 
When I ran SALVAGER, it found directories out of phase.   
I copied the most recent version to the other packs. This
worked and ITS came back up.

Later, when no one was logged in, I tried to make the system lose
by buzzing the door from a LISPM. ITS halted.

The job which got the parity error was a chaos DOOR server. It was
halted at FROBIT+1, a .SLEEP which comes immediately after an
access to the TEN11 interface.

Since the TEN/11 interface seems pretty broken, I deleted the DOOR
server from the DEVICE directory and put up a system message.

Cheers,
Chris