pcb@gator.usl.edu (Peter C. Bahrs) (02/09/90)
I used to think it was a loose connection problem that periodically caused the Netware server to crash (IBM PS/2 Model 80). So I tightened all the connectors up and no problems, I thought. Periodically the clients (IBM PS/2 Model 70's) go 'dead' and can't locate the server any longer. The server's screen will show Running MUXPRC or something similar and it says to power down the system?!@!@ Then we have to reboot the clients. What a pain... (Today after I reboot the server it said: FAT location 1041 used but file does not exist. Do you want to mark...y/n I said n??) With the above exception, the server always comes back up with no problems ??? I don't understand. Has anyone seen this? Where ,..., what have you done to rectify the situation? /*------------Thanks in advance...---------------------------------------+ | Peter C. Bahrs | | The USL-NASA Project | | Center For Advanced Computer Studies INET pcb@gator.cacs.sl.edu | | University of Southwestern Louisiana ...!uunet!dalsqnt!gator!pcb | | Lafayette, LA 70504 | +-----------------------------------------------------------------------*/
mrichey@orion.oac.uci.edu (Mike Richey) (02/11/90)
In article <867@gator.usl.edu> pcb@gator.usl.edu (Peter C. Bahrs) writes: >I used to think it was a loose connection problem that periodically >caused the Netware server to crash (IBM PS/2 Model 80). So >I tightened all the connectors up and no problems, I thought. > >Periodically the clients (IBM PS/2 Model 70's) go 'dead' and can't >locate the server any longer. The server's screen will show > Running MUXPRC >or something similar and it says to power down the system?!@!@ This is a somewhat important message. If you have a GPI or an NMI error (Running MUXPRC tipped me off) then you probably have a bad or very flaky mempry chip/board in your server. My experience has been that if the bad memory chip is located in the lower addresses of RAM, then the errors will occur often. If the failing address is higher in memory, this error will occur less frequently. Now what you need to do is either find a good set of diagnostics, like Check-It, or QAPlus and run the diagnostics on the server and identify the failing module. GPI error can be caused by a number of things. Power supplies, a NIC, system board, etc. All I can offer you is a bit of good luck. So Good luck. > FAT location 1041 used > but file does not exist. Do you want to mark...y/n >I said n??) > >With the above exception, the server always comes back up with no >problems ??? I don't understand. > >Has anyone seen this? Where ,..., what have you done to rectify the >situation? > This could be caused by faulty RAM. Netware caches reads and writes. There are easy ways of taking care of this. I really think you need get the NMI error fixed first. (The FAT error will be taken care of by Netware) Michael S. Richey mrichey@orion.oac.uci.edu University of California, Irvine Network Services
david@cwlim.CWRU.EDU (David Nerenberg) (02/11/90)
In article <867@gator.usl.edu> pcb@gator.usl.edu (Peter C. Bahrs) writes: >I used to think it was a loose connection problem that periodically >caused the Netware server to crash (IBM PS/2 Model 80). So >I tightened all the connectors up and no problems, I thought.> >Periodically the clients (IBM PS/2 Model 70's) go 'dead' and can't >locate the server any longer. The server's screen will show > Running MUXPRC >or something similar and it says to power down the system?!@!@ >Then we have to reboot the clients. What a pain...> >(Today after I reboot the server it said: > FAT location 1041 used > but file does not exist. Do you want to mark...y/n >I said n??)> >With the above exception, the server always comes back up with no >problems ??? I don't understand.> >Has anyone seen this? Where ,..., what have you done to rectify the >situation? Your problem could be caused by many possible things as, if you can find in the Novell error manual, says. In my experience, it is one of two things: 1. Your servers memory. Although IBM 80's are usually good with what type and quality of memory chips/SIMMS they have, one unreliable byte and your server goes into this process. You could try changing memory. Unfortunately, memory tests don't usually find these subtle errors because they are not always there (heat, flaw, who knows). I have found that talking the server down and making it into a DOS machine with a RAM disk taking up all of the memory and then running disk checking programs has been able to find some of these subtle errors. Both Norton DT and OPTUNE in non-cache mode have been able able to find some of my problems. 2. The other possible cause could be a dirty AC line. Without the use of an osciloscope, I know of no easy way to tell if this is the problem. If you have a UPS or a good line conditioner, this isn't the problem. My first guess would have to be the memory, that is as long as you are not running MAC VAPs! The error you received was because the server locked up when it was accessing the hard drive and the operating system FAT has an entry without a file being on the disk. The error would continue to be there on bootup of the server until the OS either uses the HOT FIX or another file uses the space on the drive intend for the file that the FAT thinks shoud be there. Basiclly, no matter what you answered to the quetion, no harm willl come to your server. Good luck.... Dave david@cwlim.ins.cwru.edu dwn@pyrite.som.cwru.edu
gribble@cica.cica.indiana.edu (gribble) (02/12/90)
I had recurring problems w/ a Model 80 server--that was eventually diagnosed as bent pins on the connector where the memory boards plug into mother board. It took months of hairpulling before I noticed the bent pins w/ a flashlight. This will give the same type of errors as the previously mentioned memory errors Actually, it was kinda of embarrassing... -- ************************************************************************ * Steve Gribble Internet: gribble@cica.cica.indiana.edu * Lead Computer Consultant swg@iumail.ucs.indiana.edu * Dept. of Sociology, Indiana University Bitnet: gribble@iubacs
chapman@acf4.NYU.EDU (Gary W. Chapman) (02/13/90)
I have had HUGE problems with a model 80 server (running NW386); we replaced the motherboard, and all is well. The bad server had '87 ROMS; the working server motherboard has '88 ROMS. I do not know if it was the ROMS or whether there was some other motherboard problem.