COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	SYSTEM MEETING							17 July 1980
C00012 ENDMK
C⊗;
SYSTEM MEETING							17 July 1980
(Some info is from after than the meeting.)

			      EQUIPMENT

DCA scanner:	Still need more replacement chips.  Will look into it.  - ROY

Ampex Disk:	Drive F has been down since the Move, but it seemed to work
		when we borrowed a card for it from another drive.  Two cards
		originally ordered for drive F arrived broken before the Move,
		and the status of their return and replacements was lost
		during and after the Move.  Will look into it.  - ROB

KL10:		Revision 10 still needed (though RPG has modified MACLSP to get
		around it).  We get occasional SBUS errors and indeterminate
		parity errors that might be fixed with later Revs than we have.

Datamedias:     All the DM 3025s arrived long ago, but with (1) poor
		quality Product Associates modems, (2) mis-positioned
		left-hand TOP and SHIFT keys, and (3) minor microprogram
		bugs.  So they will all have to be returned briefly to get
		these problems fixed.  The first one has been returned. - ME

Foonly C1:	Marginal power supply was replaced with bigger model, and new
		microcode was installed to fix the hangage problem. - ME

Ethernet        Interface successfully installed and tested in KL-10's
		front-end PDP-11.  Now awaiting high level software in WAITS.

Data Disc	Many channels are streaking a little, 3 or 4 channels are
		unusable (and detached from the system).  These problems
		may be fixable, even if disk tracks are damaged, since
		there are spare tracks available.  - ROY

Magtapes:	Drive B has been down for a while because it occasionally
		goes into "spaced-out" mode while writing.  Also, sometimes
		doesn't move forward to load point after a rewind, and now
		seems not to rewind at all.  Working on it.  - ROY

New Magtapes:	Some new (9-track) drives might be acquired, and would likely
		be put on the Ethernet to allow all systems to utilize them.
		Awaiting working E/net and analysis of drive requirements.

Dectapes:	Don't work.  Haven't for a long time.  May never work again.
		Perhaps we will get some code written in KLDCP to support
		user Dectapes on the front-end PDP-11 drives, however, that
		hardware may not be able to handle anything but PDP-11
		format tapes.

Vending Machine: Interface is up and machine works.  Now merely need
		Pony committee to keep the machine filled and food ordered.

Thermometer:	We could multiplex several different thermometer readings,
		including outside and KL vent temperatures, with not too
		much work.  Currently the wholine temperature tells the
		machine room (uncalibrated) temperature.

Audio switch:	Is up.  Speakers being installed.  Audio channel 10 (ME tuner)
		is also up now.

Disk controller, unit #0:   Won't read microcode from floppy, even with
		floppy borrowed from other controller.  Needs Ampex
		Controller wizard, possibly from Tymshare, to fix it.

Disk floppy drive, controller #0:	Has busted opto sensor?

DM dialup lines:   TTY3 still garbages substantially, possibly crosstalk.
		TTY4 sometimes gets into state where it isn't busy but
		won't answer; often enters this state by dropping carrier
		in the middle of an ongoing connection.  There are now
		10 split-speed dialups, 7-0081 is the first number for
		the entire 10-line hunt group, 0081 through 0080 (sorry,
		the phone company puts the 0 "on the right").

DCA AS cards:	All spares plus one are broken.

STATUS:		The file STATUS[1,ROB] lists various projects, mostly
		hardware, and their status as of some point.  That file
		is NOT up to date, however.


				PROGRAMS

WAITS:		The NSL "privilege" is now available to anyone who
		wants "negative service level".  Jobs with this enabled
		will only be run when no normal jobs want to be run.
		Thus ENABLE NSL is a way to ensure that a long running
		program will not interfere with interactive users,
		although NSL jobs have been observed to get 50% of the
		machine overnight.

		(This item is later than the system meeting:)  WAITS
		now supports more than the old limit of 63 jobs.  The
		current limit is arbitrarily set to 79 jobs.  At some
		point, the other old limit of 128 terminal lines will
		also be increased.

Uptime figures since the Move:

		MTBR= mean time between reloads
		MTBC= mean time between crashes

1979 Rl Null  Up   MTBR   MTBC  Sft Hrd MnS MnH Pwr   Remarks
Oct  15 18.0 43.5  24.52  52.54   2   3   6   2   2   Moved to MJH
Nov  13 84.8 51.7  13.24  21.51   0   8   1   4   0   lots of ARM-10M hang, par err
Dec  38 68.6 89.3  19.28  25.26   2  27   4   5   0   many MG overtemps, par errors

1980 Rl Null  Up   MTBR   MTBC  Sft Hrd MnS MnH Pwr   Remarks
Jan  23 32.2 97.4  28.29  34.24   3  16   1   3   0   many 1000,,0 bit par errors
Feb  23 47.8 93.8  31.37  42.45   4  12   4   2   1   many par errors
Mar  20 45.5 94.4  32.00  64.00   1   9   2   8   0   much ARM-10M trouble
Apr  16 39.3 93.8  42.27  67.64   1   9   6   0   0   DTE load work
May  22 21.6 94.1  33.83 106.34   2   5  10   5   0   alternate CTY
Jun  11 39.1 97.6  58.73 129.21   2   3   6   0   0   Ethnet debugging
Jul  26 39.3 94.0  27.09  44.01   6   8  10   0   2   Ethnet, display support
Uptime %:    ↑↑↑↑

Displays:	WAITS now supports various types of display terminals.
		Complete documentation on the display service can be
		found with the monitor command READ DISPLAY.  Other
		useful info can be found by READ NOEDIT (for EDIT-key-less
		displays) and by READ DM (for displays with EDIT keys).
 

				CONTROVERSIES

Keyboards:	A discussion ensued over the shuffling of the digits
		on the Microswitch keyboards.  A straw poll showed users
		against the change.  The autocrats considered everything
		and finally decided to undo the change.