nls-technical
[Top] [All Lists]

Re: [nls-technical] pdp-10 freezing issues

To: Jonathan Cheyer <jonathan@cheyer.biz>
Cc: nls-technical@chm.cim3.net, mrc@cac.washington.edu, Ken Harrenstien <klh@panix.com>
From: Ken Harrenstien <klh@panix.com>
Date: Sun, 21 May 2006 15:49:11 EDT
Message-id: <CMM.0.91.0.1148240951.klh@panix1.panix.com>
[cc'ing MRC in case]    (01)

> Hi Ken,
> 
> As I've been using the version of TOPS-20 that was cloned from Doug's 
> machine, I've seen a lot of cases where the PDP-10 seems to freeze, or 
> lock up, after it has been running for several hours. It typically 
> occurs without any errors displayed on the TOPS-20 console.
> 
> It is never the case that KLH10 itself has frozen. I can always return 
> to the command parser. However, the underlying PDP-10/TOPS-20 
> environment does not respond to any input from the console, and no 
> output is shown after it freezes. Also, all TOPS-20 network processes 
> (telnet and ftp servers) no longer respond at all.
> 
> When this occurs, my only option is to return to the KLH10 command 
> processor and issue a quit command.
> 
> The one error that I see fairly often, and it usually happens to be the 
> last message displayed before freezing, is:
> 
> [CONLOG] ** My spy link had been snapped **
> 
> Since I also get that message even when the emulated machine is not 
> frozen, it doesn't seem related specifically to the freezing problems, 
> but I thought I'd mention it just in case.
> 
> I don't really know how to get started trying to figure out why this 
> occurs, or even what additional information you would need to help me 
> figure out how to fix these problems. I presume there is some way using 
> KLH10 that I can give you some kind of dump that would show the state of 
> memory when it freezes?
> 
> I'm not sure what is different than Doug's setup, but his version 
> appears to run for months at a time without any freezes.
> 
> Any suggestions?
> 
> Thanks,
> 
> Jonathan    (02)

I remember encountering something like this on Doug's system.  The
workaround that I settled on was to build his system with a special
configuration that avoided the esoteric real-time-interrupt
mechanisms, which for some unknown reason didn't work well on Linux
with high performance hardware.    (03)

This isn't in the standard Makefile.  It should have been in the copy
you got from Doug, but probably would not have been in the one from
MRC.    (04)

Here it is; what diffs do you see between this and your config?    (05)

-------------------------------------------------
base-kl-bi-test:
        $(MAKER) kn10-kl $(DPROCS_KL) $(BASE_UTILS) uexbconv \
            "SRC = $(SRC)" \
            "CC = $(CC)" \
            "CFLAGS = $(CFLAGS) $(CFLAGS_AUX)" \
            "LDFLAGS = $(LDFLAGS)" \
            "LIBS = $(LIBS)" \
            "CENVFLAGS = $(CENVFLAGS)" \
            "CONFFLAGS = \
                -DKLH10_CPU_KLX=1       \
                -DKLH10_SYS_T20=1       \
                -DKLH10_EVHS_INT=1      \
                        -DKLH10_DEV_DPNI20=1 \
                        -DKLH10_DEV_DPTM03=0 \
                        -DKLH10_DEV_DPRPXX=0 \
                -DKLH10_MEM_SHARED=1    \
                -DKLH10_RTIME_OSGET=1   \
                -DKLH10_ITIME_SYNC=1    \
                -DKLH10_CTYIO_INT=0     \
                -DKLH10_APRID_SERIALNO=2 \
                -DKLH10_CLIENT=\\\"bootstrap.org\\\" \
                $(CONFFLAGS_AUX) "    (06)

-------------------------------------------------    (07)

--Ken    (08)

_________________________________________________________________
Message Archives: http://chm.cim3.net/forum/nls-technical/
Shared Files: http://chm.cim3.net/file/work/project/nls-restore/
Community Portal: http://www.computerhistory.org/  
To Post: mailto:nls-technical@chm.cim3.net
Community Wiki: http://chm.cim3.net/cgi-bin/wiki.pl?NLS_Restoration    (09)
<Prev in Thread] Current Thread [Next in Thread>