Number: 823

Date: 24-Apr-84 15':19':04

Submitter: le.pasa

Source: Bill White (Teknowledge)-Marty Raim

Subject: Incomplete sysout written by sysout to floppy

Lisp Version: Fugue.6 (March 30)

Description: '
Teknowledge loaded KS300D into an 1108 volume and tried (SYSOUT ''{FLOPPY}).  After only a few seconds into the third floppy, SYSOUT returned, although KS300 requires 3 full floppies to record the sysout.  Attempting to install the sysout with the Installation Utility produced the message "Trouble reading floppy" half way into the third floppy.  The problem is fatal, in that Teknowledge uses floppies to deliver software to their customers.  No software, no customers.  This problem happens everytime with the KS300D sysout.'
'
-----'
'
Date': 25 Apr 84 17':18 PST'
From': Roach.pa'
Subject': AR823 (Incomplete Teknowledge sysout to floppy)'
To': LISPSUPPORT, RAIM.PASA, LE.PASA'
cc': Roach.pa'
'
     With this limited evidence, I can only make a few wild guesses as to what might be happening.  "Trouble reading floppy" could just mean that the third floppy is physically bad.  If this is an actual FLOPPY problem, I might need to visit Teknowledge to get a clue on it.   But before doing that, I would like 1100SUPPORT to try to get Teknowledge to try and/or answer the following (hopefully saving me time & visits)':'
     (1) SYSOUT to a Vax instead of {FLOPPY}.  If this sysout is supposed to take up 3 floppies, the file should be near 6750 pages in length.  Use function FLOPPY.MODE to set {FLOPPY} in SYSOUT mode, then use COPYFILE to copy the sysout off the Vax to {FLOPPY}.  Is the resulting sysout on floppies ok?  (If so, this is a workaround to their problem.)'
     (2) Is the VAX sysout length exactly a multiple of 2250?  Close?  How does the length compare with the value returned by (VMEMSIZE)?'
     (3) Does (SYSOUT ''{FLOPPY}) work with something other than KS300D loaded?'
Work with an older version of KS300D loaded?  Work with an older version of Fugue?  Work with KS300D plus some extra files (to increase the size of the sysout) loaded?'
     (4) The problem is independent of which particular floppies are used to store the sysout?'
     (5) Take the 3 floppies produced by (SYSOUT ''{FLOPPY}) and do the following.'
With FLOPPY.MODE set to PILOT, insert the floppies in succession doing'
(GETFILEINFO ''{FLOPPY}lisp.sysout ''PAGELENGTH)'
(GETFILEINFO ''{FLOPPY}lisp.sysout ''HUGEPAGELENGTH)'
PAGELENGTHs should be 2250, 2250, <=2250.  HUGEPAGELENGTHs should be equal and somewhere between 4501 and 6750.'
					Kelly'
'
------'
'
Date': Thu, 26 Apr 84 09':13 PST'
From': Raim.pasa'
Subject': Re': AR823 (Incomplete Teknowledge sysout to floppy)'
In-reply-to': "Roach.pa''s message of 25 Apr 84 17':18 PST"'
To': Roach.pa'
cc': LISPSUPPORT.pa, RAIM, LE'
'
Kelly,'
'
I asked Bill White to carry out instruction (1) last night.  He reports the sysout was written correctly to the VAX.  Then he used FLOPPY.MODE to set {FLOPPY} in SYSOUT mode and used COPYFILE to copy the sysout off the Vax to {FLOPPY}.  Instead of 2 and 1/16 floppies being written, only 1 and 1/2 were written.'
'
I''ll forward the rest of your suggestions to Teknowledge.'
'
--Marty'
'
-----'
'
Date': Thu, 26 Apr 84 11':44 PST'
From': Raim.pasa'
Subject': Re': AR823 (Incomplete Teknowledge sysout to floppy)'
In-reply-to': "Roach.pa''s message of 25 Apr 84 17':18 PST"'
To': Roach.pa'
cc': LISPSUPPORT.pa, RAIM, LE'
'
Answers to more of your questions.'
'
(2) Is the VAX sysout length exactly a multiple of 2250?  Close?  How does the length compare with the value returned by (VMEMSIZE)?'
'
VAX sysout size (in disk blocks) is 6937'
VMEMSIZE returns 6628'
'
     (3) Does (SYSOUT ''{FLOPPY}) work with something other than KS300D loaded?'
Work with an older version of KS300D loaded?  Work with an older version of Fugue?  Work with KS300D plus some extra files (to increase the size of the sysout) loaded?'
'
Yes, it works OK on other sysouts. Has worked OK with older versions of fufue.  Still fails for KS300D in current version when extra files are loaded (VMEMSIZE=7126)'
'
     (4) The problem is independent of which particular floppies are used to store the sysout?'
'
Yes.  Two different sets of floppies were tried on two different 1108s.'
'
     (5) Take the 3 floppies produced by (SYSOUT ''{FLOPPY}) and do the following.'
'
Answer will be provided in next message.'
'
--Marty'
----------'
'
Date': Thu, 26 Apr 84 13':23 PST'
From': Raim.pasa'
Subject': Re': AR823 (Incomplete Teknowledge sysout to floppy)'
In-reply-to': "Roach.pa''s message of 25 Apr 84 17':18 PST"'
To': Roach.pa'
cc': LISPSUPPORT.pa, RAIM, LE'
'
     (5) Take the 3 floppies produced by (SYSOUT ''{FLOPPY}) and do the following.'
With FLOPPY.MODE set to PILOT, insert the floppies in succession doing'
(GETFILEINFO ''{FLOPPY}lisp.sysout ''PAGELENGTH)'
(GETFILEINFO ''{FLOPPY}lisp.sysout ''HUGEPAGELENGTH)'
PAGELENGTHs should be 2250, 2250, <=2250.  HUGEPAGELENGTHs should be equal and somewhere between 4501 and 6750.'
'
The answers':'
'
'
KS300.sysout  PAGELENGTH=2250, 33, 2060'
    HUGEPAGELENGTH=6644,  2283, 6560'
'
(By comparison, Lisp.sysout  PAGELENGTH=2251, 2250, 671)'
'
----------'
'
Date': 27 APR 84 18':10 PST'
From': ROACH.PA'
Subject': AR823 (Teknowledge floppy problem)'
To':   LISPSUPPORT, RAIM.PASA'
'
     Teknowledge''s problem seems to be that it is still possible for'
{FLOPPY} to misformat a floppy that has never been formatted before.'
They have actually done very little sysouting to {FLOPPY}, and when I'
asked some of the same questions in person that I asked in the'
computer mail, I got new answers.  A workaround is to do'
	(FLOPPY.FORMAT NIL NIL T)'
on any suspect or brand new floppy.  Jacques Guilemard, the person I'
worked with, was very relieved that he would be able to give a visitor'
an S1 sysout on floppies the following day.  Bill White was also '
more amiable this time around.  No difficulties with {FLOPPY} other'
than sysouting were mentioned to me.'
				Kelly'
'
-----'
'
Date': Mon, 30 Apr 84 10':53 PDT'
From': Raim.pasa'
Subject': Re': AR823 (Teknowledge floppy problem)'
In-reply-to': "ROACH.PA''s message of 27 APR 84 18':10 PST"'
To': ROACH.PA'
cc': LISPSUPPORT.PA, RAIM'
'
Kelly,'
'
Thank you for attending to this problem.  Unfortunately, it is too late to include the (FLOPPY.FORMAT NIL NIL T) caveat in the Fugue.6 documentation.'
'
--Marty'
'


Workaround: do a (FLOPPY.FORMAT NIL NIL T) on new floppies before doing the sysout

Test Case:  On-site, Bill White''s Dandelion

Edit-By: Sannella.PA

Edit-Date: 17-Aug-84 11':03':32

Attn: 

Assigned To: Roach.pa

In/By: Harmony

Disposition: [4/26/84 tl': more info are appended to original message by lmm and tl]'
[DATE': 23-JUL-84 ROACH This was marked Fixed 3-MAY-84, but still attn''d to me.  ATTN _ LISPSUPPORT]

System: Operating System

Subsystem: DLion Floppy

Machine: 1108

Disk: 

Microcode Version: 

Memory Size: 

File Server: 

Server Software Version: 

Difficulty: Hard

Frequency: Intermittent

Impact: Fatal

Priority: Absolutely

Status: Closed

Problem Type: Bug

Source Files: