 |
Given that there are literally millions of computers that run CentOS 7 without issues, the chances that you have found a bug that triggers a system crash and that it affects only you are extremely low. Please run a comprehensive test of your hardware because it is much more probable that you have a hardware issue such as overheating, a defective storage drive ( which in turns leads to bad code or data being loaded by the OS ) or a bad memory chip. Incidentally you are also using a version of the kernel which is slightly out of date. I doubt that this is your culprit but a yum update should not hurt. You should also pay a visit to https://support.huawei.com/enterprise/en/server/rh2288-v3-pid-9901877/software/ because you do not seem to run the latest BIOS and maybe there are newer versions of interest for you. |
 |
We reported this server to the manufacturer for repair several times in 2022, and also replaced a memory bar. Later, it still crashed and restarted frequently, but the manufacturer later checked that there was no hardware problem. I have also upgraded the system kernel several times. I upgraded the kernel yesterday to the latest version 3.10.0-1160.83.1.el7.x86_64, we will continue to observe. The previous crash date is referenced below： ls /var/crash/ 127.0.0.1-2022-09-23-19:08:30 127.0.0.1-2022-09-24-12:58:48 127.0.0.1-2022-10-19-00:09:39 127.0.0.1-2022-12-21-04:26:01 127.0.0.1-2022-09-24-07:10:10 127.0.0.1-2022-10-13-15:13:39 127.0.0.1-2022-12-08-07:00:54 127.0.0.1-2023-03-03-01:29:08 |
|
|
Many years ago I had a period when I was writing very manyFedora DVDs monthly, shipping them to people who could not afford to download them. Unfortunately roughly half of the disks I wrote turned out to be defective. Nothing else seemed to be wrong with that computer so I suspected the disks and obviously the DVD writer. After replacing disk brand, DVD unit, cables and IDE port I did a full and thorough hardware check of the whole system. Turned out that a memory module was defective but memtest+ would report it as perfectly fine as long as the computer was configured to run in dual channel mode. The moment after I switched to single channel, memtest+ failed. Replaced the faulty module and lived happily ever after. As another story, the very system I am typing from used to blow up NVME SSDs ( they just became unreachable as if nothing was connected to the NVME port of the motherboard ) after a random period of usage ( between a week and 3 months ). Kingston were nice and replaced them under warranty but after the 4th replacement they suggested to attempt a BIOS update. Guess what ? Did that and the system works perfectly fine since Jan 2021. What I mean with the above stories is that I still bet on a hardware issue even if your supplier was not able to identify any issue. There are tons of things that can go wrong, from power supply, bad contact between cooler and processor, memory.. even a defective processor that acts up in very specific conditions ( yep, I've seen that too, I have an AMD Duron that seems perfectly fine unless you run a very very very specific test ). The chances that in the whole world you are the only one affected by a constant crash of CentOS 7 are close to none. |
|
|
Ok, now that I have your specific confirmation reply, I will continue to try to check the hardware. No matter what the result is, I will consider replacing all the memory first, etc. If the fault is more frequent, I will even consider replacing the whole server. |
- Anonymous
