I have a problem with my ubuntu system. During normal work i receive logs on screen presented bellow. I'm also not able to log in via ssh after sometime of system working. Also software working on it due to errors presented is not working in a correct way. Can anyone help me diagnose what might be a problem?
Welcome!
That is a kernel panic, and it marks the UKSM kernel module "tainted".
Either it's a bug in the kernel or the RAM is faulty.
@mm211 , this link may help tainted kernels
FWIW, my money is on a RAM problem. I suggest that you download a memory check ISO, burn it, boot from it and see what result you get. If the fault is at a high address then you wouldn't necessarily see a crash until the system is working hard.
I performed test with memtest86 and i receive: [ECC Errors] for test 5,6,8,9,10. There are not many of them, final test result shows "PASS" each time, however each time i run a test I receive those errors and its only in one RAM slot (except one time out of ~10 tests). I used different RAM stick and no matter witch slot i put them error persist in "Channel-Slot: 5-0" only. I switched also CPU between two slots and i got same results. What would you guys suggest? Is this a cause of my initial problem or it shouldnt have any affect on performance of my server? are those ram errors caused by motherboard, bios version or something else? Because it seems as its not RAM problem itself.
Is your "box" inside a hypervisor like kvm or qemu?
Then there might be a bug in this software, too.
I think your test shows corrected errors. No real harm, but something is not good there.
Likely no impact on your crash problem.
Its bare metal with ubuntu and software for virtualization, however its ubuntu (native system) that report errors not any vm on top of it.
Hmmmm......I've seen this kind of thing happen if a 32bit memtest is used on a 64bitbox. If the mobo is 64bit make sure you use the correct diag.
Otherwise, try and test half the memory modules at a time if the mobo will let you.
Very likely you are identifying the problem here
By any chance are you running a SuperMicro board such as the X10DRL-i? I am getting the exact same error message pop up (Also on Channel-Slot: 5-0) when running MemTest with 256GB DDR4 2400MHz Registered ECC ram. Were you by any chance able to pinpoint the cause of these errors?