LVM crashes on debian 9. Root directory on LVM, system unwritable, fsck helps for a few hours, then relapses. How to be?

Question

Debian

Background.
A "wonderful" "specialist" came to clean and prevent the server from the company from which we receive outsourced sysadmin services.
Server under ESXi 6.5. The accountant and the director - girls, having turned off the VM with 1C, reported that the server was turned off. This "specialist" cuts down a running server under a hypervisor from the network, disassembles it, removes the screws, disconnects them from the RAID controller, and cleans the whole thing. Collects - of course nothing works, except for what was not on the RAID disks.
Then the following message appears on the debian 9 machine:

I put the disk for checking manually, correct all errors, debain starts up and works fine for 4-6 hours, and then everything repeats again. The system loses the availability of the disk for recording, after rebooting what is in the screenshot. When rebooting after fsck, only messages remain:

Changing use_lvmetad from 1 to 0 in /etc/lvm/lvm.conf helps, but the machine and sites on it start to slow down fiercely (((

How can I fix it humanly?

Answer 1

Picked up the problem.
ESXi is on an Intel S3710 200Gb SSD drive, and for some reason it just loses the ability to write after a while (((
In the ESXi system it writes 0B Volume, 0V Busy 0V free.
I'm testing it under Windows Crystalmark, everything is ok!
ESXi only yesterday I reinstalled, refreshed the release, while doing a full format.Maybe
because of the PSU, there is a drawdown on the 5V line to 4.83-4.85V
In general, a mystery.