Conference bulova::decw_jan-89_to_nov-90

1607.0. "Spacewar killed my manager!" by LARK::CRAIG (Western Area SW Support) Thu Oct 19 1989 20:50

    
    When I tried to exit from DECW$SW (Spacewar), all DECW processes
    died. Obviously this isn't a crucial application, but I'm wondering
    what may have happened. It appears to be the fault of DECW$SW
    (2db821c is an xlib io error) but I don't understand the reference to
    an unexpected message from DECNET ACP in the server error log.
    
    Is there a "Guide to Cryptic Xlib Fatal Error Message" available??
    
    Thanks,
    
    Craig
    --------------------------------------------------------------------
    Platform: VS2000 6MB VMS V5.2 DECW V2
    
    Contents of decw$sm.log:
    
%SET-I-NEWLIMS, new working set:  Limit = 200  Quota = 200  Extent = 2048
XIO:  fatal IO error 65535  on X server "LARK::0.0"
      after 1948337 requests (1948297 known processed) with 0 events remaining.
Session Error: %NONAME-F-NOMSG, Message number 02DB821C

XIO:  fatal IO error 65535  on X server "LARK::0.0"
      after 1948338 requests (1948297 known processed) with 0 events remaining.
%NONAME-F-NOMSG, Message number 02DB821C
    
    
    Contents of decw$server_0_error.log:
    
 6-OCT-1989 12:50:10.4 Hello, this is the X server
Dixmain address=1440c
Now attach all known txport images
in SetFontPath
out SetFontPath
GPX color/monochrome support loaded
gpx$InitOutput address=109188
Connection Prefix: len == 42
 6-OCT-1989 12:52:04.6 Now I call scheduler/dispatcher
 6-OCT-1989 12:52:16.2 Connection 9a100 is closed by Txport
 6-OCT-1989 12:55:19.9 Connection 9a138 is closed by Txport
 6-OCT-1989 14:29:43.4 Connection a0ef8 is closed by Txport
10-OCT-1989 12:28:32.0 Connection 9dda8 is closed by Txport
10-OCT-1989 12:28:42.9 Connection 9c228 is closed by Txport
10-OCT-1989 12:30:29.5 Connection 9c478 is closed by Txport
10-OCT-1989 13:39:06.5 Connection 9c478 is closed by Txport
10-OCT-1989 15:44:18.9 Connection 9c4b0 is closed by Txport
10-OCT-1989 15:44:45.8 Connection 9c478 is closed by Txport
16-OCT-1989 12:13:29.4 MEMMGR/XREALLOC - zero amount is requested
16-OCT-1989 12:13:31.0 MEMMGR/XREALLOC - zero amount is requested
16-OCT-1989 12:13:31.6 MEMMGR/XREALLOC - zero amount is requested
16-OCT-1989 12:13:37.2 MEMMGR/XREALLOC - zero amount is requested
16-OCT-1989 12:15:20.9 Connection 9c4e8 is closed by Txport
16-OCT-1989 12:21:02.6 Connection 9c4e8 is closed by Txport
16-OCT-1989 12:21:31.1 Connection 9c4e8 is closed by Txport
16-OCT-1989 13:37:42.0 Connection 9c4e8 is closed by Txport
16-OCT-1989 14:32:08.1 Connection 9c4e8 is closed by Txport
18-OCT-1989 13:30:53.6 Connection 9c4e8 is closed by Txport
18-OCT-1989 14:16:04.1 MEMMGR/XREALLOC - zero amount is requested
18-OCT-1989 14:16:04.9 MEMMGR/XREALLOC - zero amount is requested
18-OCT-1989 14:16:06.2 MEMMGR/XREALLOC - zero amount is requested
18-OCT-1989 14:16:11.1 MEMMGR/XREALLOC - zero amount is requested
19-OCT-1989 15:53:31.5 Connection 9c4e8 is closed by Txport
19-OCT-1989 15:54:12.9 Connection 9c4e8 is closed by Txport
19-OCT-1989 15:57:57.3 Using extra todo packet pool...
19-OCT-1989 15:59:07.8 RegionReset - BAD BOX 1001,42 1001,84
19-OCT-1989 15:59:13.7 RegionReset - BAD BOX 1001,42 1001,84
19-OCT-1989 16:01:52.7 RegionReset - BAD BOX 1001,42 1001,84
19-OCT-1989 16:03:41.5 RegionReset - BAD BOX 1001,42 1001,84
19-OCT-1989 16:05:42.7 RegionReset - BAD BOX 1001,42 1001,84
19-OCT-1989 16:11:32.0 RegionReset - BAD BOX 1001,42 1001,84
19-OCT-1989 16:11:38.4 Connection 9c4e8 is closed by Txport
19-OCT-1989 16:11:41.1 %SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=0000000E,

Unrecoverable server internal error(error code = 12) found, terminating all connections.
Exception Call stack dump follows: 
	904f5
	f395
	104cc3
	1059c3
	1b024
	eada
	1bf3e
	1394f
	34515
	1298f
	12481
	14705
	41a
	80173ed3
	80173e84
********** marking the end of call stack dump **********
********************************************************
%SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=200BFFFE, PC=0000F395, PSL=03C00

  Improperly handled condition, image exit forced.

	Signal arguments	      Stack contents

	Number = 00000005		 7FF9AE90
	Name   = 0000000C		 00104EA8
		 00000001		 0028AEB0
		 200BFFFE		 00000000
		 0000F395		 21FC0000
		 03C00000		 7FF9AEB8
					 7FF9AEA0
					 00104CC3
					 0028AE50
					 0028AEB0

	Register dump

	R0 = 000AAD00  R1 = 200BFFF8  R2 = 200C0000  R3 = 0028AEB0
	R4 = 00051FC0  R5 = 7FF9AF5C  R6 = 00000041  R7 = 0004D598
	R8 = 0000F6B8  R9 = 00000000  R10= 00000040  R11= 0004EB2C
	AP = 7FF9AE18  FP = 7FF9ADD8  SP = 7FF9AE54  PC = 0000F395
	PSL= 03C00000

%DECW-F-UNEXPECTED_MESS, unexpected message from DECnet ACP (message code 0)
    

This may be hard to find, but the server error log gives you a clue:
"Unrecoverable server internal error(error code = 12) found, ...",
and "%SYSTEM-F_ACCVIO, access violation...".

What this means is that the DECwindows server ACCVIO'ed.  When the
server dies, all its DECnet connections will terminate, and that is
why all your applications will receive DECnet errors and stop.

User applications should never be able to crash the server.  Therefore,
you should QAR this problem in the V5 QAR database on TRIFID::.

You can ignore the "unexpected message from DECnet ACP", as the server
was dead at that point anyway.

"The Guide to Cryptic Xlib Fatal Error Message" is available as note
1532 in this very conference!

    
    Thanks, I had noticed the ACCVIO. I was curious about the DECNET ACP
    message because default transport is local and ignore_decnet is YES.
    This problem has been QAR'd (3690) in DECWINDOWS-IFT.
    
    Craig	

    
    
    Sometimes when I am running DwtEtris, and I have a DECTERM window or
    two on the screen and I start switching back and forth between windows
    and the game, I start to hear my system working hard.  This is usually
    followed by my GPX hanging.  
    
    I log on from somewhere else and do a SHOW SYSTEM.  There are three 
    unnamed processes in RWMBX (something close to that) status.  If I try
    to stop them, it says non-existent.  Is there a way to resolve this
    hung situation with out shutting down and booting again?
    
    Lisa

I've had a series of hangs thoughout my time with DECwindows.

AUTOGEN with FEEDBACK has nearly always solved my problems (except for the time
that the difficulty was due to a bug in the DECwindows server).

After your system has been up for a while, and is "working hard", try an AUTOGEN
(with FEEDBACK), and see if that changes or eliminates the problem.

	-- Jeff

    
    Yes I did run AUTOGEN, as specified by a SYSTEMs person, after it hung
    the first time.  It has since hung again with the same situation
    running.  Should I AUTOGEN every time it hangs and reboot?
    
    I am a satellite on a LAVC, FWIW.
    
    lisa

Sounds like your MODPARAMS.DAT has too much in it.  If the MODPARAMS.DAT is
left over from V4.n then it will contain far too many specifications.  You
want to reduce the contents to the absolute minimum and then use AUTOGEN
and make sure you use FEEDBACK.

Dick

    Re: .0,.1
    
    You should also upgrade to the SSB version of VMS V5.3 so that you get
    the final version of DECwindows V2.  I fixed the bug that was causing
    the "Unexpected message from DECnet..." to appear.  I also fixed a
    timing bug which would cause the server to crash during client
    disconnect (very infrequently).
    
    James