January 4, 20215 yr Dear all, Happy new year, may you and your loved ones safe and happy for 2021! I recently upgraded my server to Ryzen 3000 with 32G 3600 CL16 memory following the great advice from the community. However, since I set up the system I have experienced twice random reboots (Machine Check Events detected as suggested by Fix Common Problems). The latest one happened just after I shut down my Windows 10 vm. I have installed the mcelog since the first reboot but I cannot seem to find the mcelog in the latest diagnostics. Can anyone have a look and see if the unclean reboot is related to the vm or something more serious has happened? Some info for the vm: This vm uses a RX580 Pulse primary GPU passthrough. It was booted from a dedicated Samsung 970 Evo nvme drive migrated from an existing vdisk to the nvme for dual boot (following Spaceinvader One's videos). It took me some time to set up the GPU passthrough (using the Q42 machine solved the problem in the end) but I have to use the 170310 ROM from TechPowerup. Thank you very much for your help in advance. tower-diagnostics-20210104-1442.zip
January 4, 20215 yr Community Expert https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=819173
January 4, 20215 yr Author 3 minutes ago, JorgeB said: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=819173 Thanks for your reply. I am aware of the memory speed limit of 3200. However, it appears that the memory instability tends to lock up the server, rather than causing reboot. Have you got a chance to look the diagnostics and see if there is any problem? Thanks.
January 4, 20215 yr Community Expert Diags after rebooting are not much help, first thing to do is to stop overclocking the RAM, then test if it still crashes, also make sure power supply idle controller is correctly set, as per the link.
January 6, 20215 yr Author On 1/4/2021 at 3:31 PM, JorgeB said: Diags after rebooting are not much help, first thing to do is to stop overclocking the RAM, then test if it still crashes, also make sure power supply idle controller is correctly set, as per the link. Thanks, that makes sense. I have tuned down the memory to 3200, but turns out it was because of a conflict of USB controllers. I passed through the upstream "Non-Essential Instrumentation [1300]: Advanced Micro Devices, Inc. [AMD] Starship/Matisse Reserved SPP". But as soon as I entered Windows the whole vm crashed. I will now test the stability of 3600 Thanks for your advice!
Archived
This topic is now archived and is closed to further replies.