How to diagnose iron?

Question

Debian

From time to time the system hangs tightly (debian). You need to understand the reason. The hard drive has nothing to do with it, I installed a new one, at first I thought that it was a problem. I ran the RAM through Memtest86 once, no errors were found. The temperature seems to be in order, there is no big load. I also did stress testing . I want to do a complete hardware diagnostic from some kind of livecd (for example, Inquisitor).
PS
Most likely solved my problem. It was all about the Chinese KLLISRE RAM. In the BIOS (e), I lowered the frequency from auto (1600 GHz) to 1333. Most likely, the Chinese soldered different chips (I tried to see it, I didn’t see it). The RAM got hot and the computer froze. Multiple testing of memtestom86 and 86(+) installed from the repository did not give anything. Did 6 laps in 12 hours. But from a flash drive with winpe 10-8 installed on it, sergei strelec was revealed through dos.
Thanks to all!

Answer 1

system (debian). You need to understand the reason.

Most likely just in
Here is a good list of boot disks with diagnostic utilities: ruterk.com/vse-obrazy-avarijno-zagruzochnyx-cd-dvd-usb-diskov
But I still wonder that the problem is in Debian.

Answer 2

For hardware:
Memtest must be run many times around. It happens that it doesn’t pour on the 4th test, but on the 5th and 6th it pours (it warms up apparently).
Blow everything from the compressor from dust (only dismantle the fans, you can tear off the impeller). Under the chips with legs at the bottom (when they are not visible at the edges), you can first plop the cleaner (white spirit, acetone) and then blow.
Check all capacitors, replace if necessary.
Wipe with an eraser all the contacts of the RAM, the processor (if not the needles but the panel); Wipe the contacts of the processor (motherboard) of the pin with a cleansing agent with a soft toothbrush, very gently so as not to bend and the RAM connectors could form a fatty film. Clean the sata and piania connectors of the mother and hard drives with an elastic band. Twist the board from the hard drives, remove oxides from the contacts in the same way.
Replace sata cables (they are not cleaned well).
Naturally thermal paste, so that there are no air pockets. I lay down a slide and evenly press the radiator to spread.
Do not forget about the BP, major cleaning. Very important.
Lubricate all the coolers by flushing with cleaners beforehand from the bushings (bearings) the remains of the old oil.
Further measurements of the state of the keys on the mother and other components must be made already in the workshop with desoldering, a tester, a cell and a laboratory power supply.
You can simply remove the video card and cling to the remote.
This all refers to what can be done at home with respect to iron. It is necessary to speak about software separately and IMHO after major purges.
I did not specifically use cool buzzwords.
We should go and not with checkers (anegdote).

Answer 3

what's in the logs?

Answer 4

Try to use a different kernel, for example, older or newer than the current one. Maybe some proprietary drivers fail.
My ubuntu 18.04 worked disgustingly on my PC, as a result I returned to Mint, the flight was normal.

Answer 5

recent experience:
barebones from the shuttle hovered (7 pieces in total, 24x7 operation mode), without declaring anything in the logs, without apparent reasons and patterns.
I sinned on the SSD model, they were equipped with Samsung 850-pro-sata, which seemed to have a tendency to fall asleep and not wake up.
They changed the SSD to Intel with power saving disabled, it became less common but not by much.
Hardware firmware updates also did not lead to anything.
It was cured only by completely disabling the C1 / C3 power saving modes of the CPU in the shuttle BIOS, since this feature appeared in the new version.