Today I was going to upgrade from 24GB to 48GB system memory using 6x4gb dimms.  I shut down as normal, installed the memory, and started back up.  I got a KERNEL PANIC: not syncing VFS: unable to mount root fs

 

I searched this on the forum and it seems like the fix is to try reinstalling unRAID on the flash, so I did that.  Didn't work.  Tried swapping to a new flash, didn't work.  Tried unplugging the cache drives.  Tried unplugging all new memory.  Tried unplugging all dimms except 1 stick per CPU.  Nothing helps.  Every once in a while I will get a freeze on a string of hex's that says --end trace-- at the end, but same result.

 

I am currently running memtest86 with just the single dimm on each CPU in case somehow one dimm slot or CPU died all of a sudden.  Any ideas??

 

Pics of kernel panic at bootup:

https://imgur.com/a/tw1AnfS

 

 

Edited April 19, 20197 yr by Benson

57 minutes ago, Benson said:

 

 

I tried that, and worked immediately after.  To test if it was really fixed, I did a reboot once again.  Back to the same error (but now slightly different). 

 

 

I ran memtest for a 4th time, and within the first second of testing it was all red -- bad CPU maybe?  Would a dying CPU boot up sometimes and fail others?  The system is Rock solid when up and running, that's why I'm so hesistant to think that it's bad hardware.

From screen dump, it is install 40GB memory ( not 48 or org. 24GB ) and running in tripper channel instead 4 channel. Could you show what is current stick config.

 

DDR3-3100 it seems display wrong, btw pls check this. Pls make sure all slot and sticks were good first , then may need slowdown memory clock rate or tune config. in BIOS.

Edited April 19, 20197 yr by Benson

1 hour ago, thrroow said:

that's why I'm so hesistant to think that it's bad hardware.

With errors during memtest there's no doubt about that.

Double check that the memory manufacturer and part number are on the MB manufacturer's 'approved list'.  Some MB's are picky about memory.  (Increasing the number of memory modules can cause problems where none existed before because the line drivers are not able to maintain the required waveforms to properly access the memory.)

6 hours ago, Benson said:

From screen dump, it is install 40GB memory ( not 48 or org. 24GB ) and running in tripper channel instead 4 channel. Could you show what is current stick config.

 

DDR3-3100 it seems display wrong, btw pls check this. Pls make sure all slot and sticks were good first , then may need slowdown memory clock rate or tune config. in BIOS.

The 40gb was intentional, I was removing 2 sticks at a time and trying different dimms to see if that helped at all.  Most of the testing I've done is with 1 stick per CPU that I know were working.

 

I will check the speed, thanks.