(SOLVED) Ubuntu Server 20.04 freezes twice a week

cn flag

My ubuntu server (20.04.3 LTS, headless) freezes almost twice a week.

The first time, the last message in journalctl was

systemd: Finished Message of the Day.

and the second time, one from jellyfin. The weird thing is (I think), the computer isn't really crashing. It freezes and the fans keep spinning. To get it working again, I always have to turn it off by holding down the power button and then turn it back on. Maybe that's obvious, but I have also seen a lot of post where the user was still able to ssh into the machine. This isn't the case. Even though, the light on my LAN switch is blinking (on) the router/modem says it's offline.

I'm using

  • CPU: Intel Core i5-6500 (4)
  • Ram: 16GB with Swap: 4GB

Does anyone know how to troubleshoot this accordingly?


Output of: $ journaclctl -b -1 -e


Should be the dmesg log file of the crash:


Ok, so I was able to evade any freezes over the past days by just rebooting once a day. This proves it's not some kind of thermal issue. Any further suggestions?


Thank you @thomas-ward for the tip. I have applied the recommended changes from this command: Unfortunately, I couldn't find the old log from the freeze-up anymore. I will reboot once more to make the changes take effect and then wait for the next freeze-up.


Uptime: 6 days, 3 hours, 53 mins

Looks like that was the fix. Thank you very much @thomas-ward. If you'd like to submit the answer, I can tag it as the solution.


Yikes. Just crashed again after doing a captcha in my jdownloader docker. Although it didn't get resolved fully @thomas-ward's recommendation did improve the situation a lot.


After a freeze during reboot, I have decided to blame these problems on hardware. I replaced the harddrive with an ssd, switched to debian (unrelated) and updated the bios. Will add coming information about this in the future. Could take some time though.


Uptime: 9 days, 13 hours, 37 mins

I think it's pretty save to say the problem is fixed now.

Notes for users with similar issues: try to 'disable' 5% of your RAM as explained in EDIT 4. If that doesn't work, try switching SATA ports on your mainboard or get a new SATA cable. In my case, logs were impossible to show up due to the HDD straight up failing/disconnecting while the computer is running.

admin avatar
us flag
Before the server crashes check `dmesg` logs. A general maintainance would be good. It may also likely be a hardware problem see if your processor is overheating and needs repasting. Do a memtest to check if your memory is still ok.
waltinator avatar
it flag
After a "sudden shutdown", aka "system crash", and reboot, the terminal command `sudo journalctl -b -1 -e` will show you the end of the previous boot's log's. If there is no hint there, suspect power/ overheating. Read `man journalctl` to see how to extract information from the logs. You can find how I make use of `journalctl` easier at ``.
cn flag
@admin Hm, thank you for the suggestions, but I dont think it's an overheating issue. I have been using cpu encoding for three days straight, but the crash was almost 10hours after that. Also, how would I go about logging dmesg before the crash? I found a log file of dmesg but I can't see any problems. (EDIT 2)
ru flag
Check syslog for any output. It might be RAM exhaustion which means you're doing things that your system just doesn't have the resources for, which can lead to these types of freezes.

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.