Unstable System - Process crashing on Win and Linux

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • phillpower2
    PCHF Administrator
    • Sep 2016
    • 15209

    #16
    Sounds like a plan and use your preferred choice of Linux (y)

    Going offline soon and not back until tomorrow so please do any testing at your convenience.

    Comment

    • DrunkBlood
      PCHF Member
      • Mar 2020
      • 119

      #17
      For the time I had today I did 3 configs with the bionicpuppy64 distr of Puppy Linux
      1. Stick 1 | Slot 1
      2. Stick 2 | Slot 2
      3. Stick 1 & 2 | Slot 1 & 2

      Over all configs I looked through syslog and kernel log
      In common:
      Overall the system is more stable that Windows. I do no have frequent blackscreens and can operate basic programms. Only the browser breaks down and I can not use it.
      In the logs I found many segfaults and after some time it logs some machine check exceptions from the CPU ie. Hardware failure. Some segfault are from dynamicly linked libs and I have not checked what programm crashes. I could play some games and run paint.
      From the system I can copy the minidump from the crash of windows(setup phase) if you need it.

      Other:
      In config 3, booth sticks, my Display Server(Xorg) crashed because of a segfault.

      All in all it seems more like an Hardware Error from memory. I checked once the Taskmgr and it said that the system uses about 300MB of RAM.

      To note: I have 2 other RAM sticks who were replaced by my current ones when the issue first surfaced.

      Comment

      • phillpower2
        PCHF Administrator
        • Sep 2016
        • 15209

        #18
        Sorry but you need to test each stick individually in each of the four slots as it is the only way that you can obtain conclusive results, a stick of RAM can work fine in three out of four slots but for some reason it causes problems when the same stick of RAM is inserted into a particular slot, this is why testing RAM with Memtest takes so long.
        Originally posted by DrunkBlood
        All in all it seems more like an Hardware Error from memory. I checked once the Taskmgr and it said that the system uses about 300MB of RAM.
        RAM or one or more of the DIMM slots.

        Test each stick individually in each of the four slots and see what happens, from there we can move onto Memtest if needs be, don`t want to put up a wall of text so will hold off posting the Memtest info for now.

        Comment

        • DrunkBlood
          PCHF Member
          • Mar 2020
          • 119

          #19
          So should I look for the system booting, no mce erorrs or/and no segfaults when testing?

          Comment

          • phillpower2
            PCHF Administrator
            • Sep 2016
            • 15209

            #20
            We are testing the RAM atm so any hard shut downs are what we are looking for.

            Comment

            • DrunkBlood
              PCHF Member
              • Mar 2020
              • 119

              #21
              Okay. Going to report back when I tested the other 6 configs.

              Comment

              • DrunkBlood
                PCHF Member
                • Mar 2020
                • 119

                #22
                Progress report:

                Done all configurations for booth sticks ie:
                1 in 1 2 3 4
                2 in 1 2 3 4
                I had no crashed and let the system run for about a min

                Now I will do the 2 out of 4 configs ie:
                1 2 done
                TODO
                13,14,21,23,24,31,32,32,41,42,43
                Before I start and waste maybe time. Is 1 min wait enough for a config after a succes. boot? Or do I need to wait like 5min?

                Comment

                • DrunkBlood
                  PCHF Member
                  • Mar 2020
                  • 119

                  #23
                  Today I had enough time to test all 2 out of 4 configurations.
                  In total I have tested 1 and 2 in all 4 slots (previous post)
                  And now I got the 2 out of 4 tests (boot up ram only ie no install and wait ~3min)
                  Stick1 Stick2
                  1|2 good
                  1|3 good
                  1|4 good
                  2|1 good
                  2|3 good
                  2|4 good
                  3|1 good
                  3|2 good
                  3|4 good
                  4|1 display server X-org crashed twice but I had no system crash
                  4|2 good
                  4|3 good

                  Comment

                  • phillpower2
                    PCHF Administrator
                    • Sep 2016
                    • 15209

                    #24
                    Can I ask that you only post back when you have finished all testing and so have all of the test results, I never got any notification of your replies #22 and #23 as you were replying to your own reply #21

                    One minute is nowhere near long enough I`m afraid and what you ideally should have been doing was to use the computer as you normally would albeit with only the one stick of RAM inserted, below is my canned info for properly testing the RAM;

                    How to physically test your RAM.

                    [COLOR=rgb(184, 49, 47)]The following checks require the computer case to be opened so take the following safety precautions 1st, disconnect the power cord from the wall socket, press the case power button for twenty seconds or so to get rid of any residual charge in the system, take anti static precautions before touching anything inside, you can do this by touching a bare metal part of the case or PSU or if a notebook by touching a household radiator or associated copper feed pipe.

                    Have a pencil and notepad to hand.

                    Remove each stick of RAM and blow out the memory slots.

                    Insert the first stick of RAM in memory slot one, reconnect the power and in the case of a desktop computer the video to screen cable.

                    Power up the computer and see how it goes.

                    Make a note of the results.

                    Repeat the procedure until the first stick of RAM has been tested in each memory slot and the results written down.

                    Remove the first stick of RAM and put it to one side on top of a piece of paper with the number one on it for identification purposes.

                    Repeat the procedure with all RAM until each stick has been tested in all slots, the results written down and the sticks identified numerically.

                    Now the really long winded part;

                    If no individual stick of memory or slot on the MB has an obvious problem all of the above needs to be repeated but this time by running Memtest 86 for each stick of RAM and in each slot on the MB and for at least 4/5 passes, for some reason that many of us cannot explain a stick of RAM may work in one slot on a MB but not in another.[/COLOR]

                    Comment

                    • DrunkBlood
                      PCHF Member
                      • Mar 2020
                      • 119

                      #25
                      As you can see in the pdf I followed the steps you gave me. In the first part of the test I ran the system for 30 min at a minimum and had mostly process crashes and could not use the browser. Only when stick 1 was in slot 2 had I a kernel panic booting up.
                      In the second part I did the standard settings for memtest86 for each config taking about 2:30h with 4 passes. Overall I had 0 errors.

                      Comment

                      • phillpower2
                        PCHF Administrator
                        • Sep 2016
                        • 15209

                        #26
                        Originally posted by DrunkBlood
                        Only when stick 1 was in slot 2 had I a kernel panic booting up.
                        Do you have a crash dmp for the above.

                        Did you have any problems when not connected to the internet.

                        Can you post a new MTB log for us.

                        Don`t know about you but I would consider four passes of Memtest to pretty much prove that the RAM and memory slots are ok.

                        Comment

                        • DrunkBlood
                          PCHF Member
                          • Mar 2020
                          • 119

                          #27
                          I think I do not have a dump for that since it is a live distro.

                          Yes I still got process crashes without my LAN connection.

                          I would like to provide a MTB log. But I can only provide that if I reinstall win since I do not have a user set up atm and so on (described in my earlier post).
                          So I would download an iso (gus link) and burn it on an usb with rufus and try to install again hoping that it won’t corrupt win again, try to install chipset drivers and then run MTB if that is ok by you.

                          Checklist:
                          RAM : ok
                          MB RAM-slots: ok
                          PSU : (new) ok
                          GPU: ?
                          drives : ?
                          MB other : ?
                          something I might have forgotten: ?

                          Comment

                          • phillpower2
                            PCHF Administrator
                            • Sep 2016
                            • 15209

                            #28
                            Wasn`t aware that Windows was not installed tbh.

                            Clean install of Windows, the chipset drivers and nothing else to begin with sounds like the way to go, no rush here as I have an early start so calling it a day now.

                            Comment

                            • DrunkBlood
                              PCHF Member
                              • Mar 2020
                              • 119

                              #29
                              Today I burned the iso as I said and started the install process. All files were copied and then it needed to restart. After the restart it showed me the setup ( which location I am from). Before I could enter anything it crashed and rebooted.Then it promted me to connect to the internet. Until that was the same as the previous install I mentioned here. But with this one I connected via cable and clicked next. Soon after that it crashed again. After some crashes I got an bluescreen CRITICAL_PROCESS_DIED. After some more reboots windows ran the recovery part. I got into troubleshoot and tried to get a cmd. But I was prompted by and passoword request of Administrator. After testing admin and root I got in with nothing as pwd. After getting in the cmd I tried to cd C:\ but the keyboard-layout was US and I do not know anything about that. So I shutdown and changed the boot order to the live linux. In that I was able to copy the 2 crash dumps which I attached here.

                              So now I have on the SSD the files of Windows 10 1909 but I can not get an MTB log.

                              Comment

                              • phillpower2
                                PCHF Administrator
                                • Sep 2016
                                • 15209

                                #30
                                Both crashes list PspCatchCriticalBreak+115 and ntkernel but no drivers involved, we have been focusing on the RAM when perhaps we should have looked at the storage devices.

                                Being honest I did not know what PspCatchCriticalBreak+115 was so used Google, fast boot and a flaky storage device was all that I found.

                                If fast boot is enabled, disable it please and see if that helps.

                                Pain in the proverbial I know but do you have another drive that you could install Windows to just for testing purposes, the SSD would need to be removed from the MB when you do this.

                                Comment

                                Working...