recurring hard drive corruption all of a sudden on two different computers!

[rebeltaz]

Before assuming it is the kernel, I'd rule out bad hardware issues. Besides having to systems with the same issue, I'd look at the HDD, SSD, and cables to those devices. A bad update _could_ have happened, which would lead to something upstream - like the router/switch introducing corrupted files.

I'm running 18.04 and haven't seen kernel issues that weren't traced back to hardware problems. It is the HWE kernel:

$ uname -r
5.4.0-80-generic

https://wiki.ubuntu.com/Kernel/LTSEn..._Bionic_Beaver

After you rule out the other hardware issues, only then would I look at software or corruption or kernel issues. Do the easy things first. SMART data should be clear if the storage is having issues. Run long tests, then do a report.

If your storage has any odd components, say Intel Optane (or whatever they call it), then that would be important to note and search for issues around.

SMART reports no issues. I'll let it do a test over night just to be sure, but...

I don't know what Optane is, and I probably couldn't afford anything "odd". The hard drive in this laptop is just a Samsung 850 EVO SSD. I reckon that's normal enough?

If it were JUST my computer throwing the same errors, I'd be more open to hardware issues (not that I will rule that out, though) but with two different computers both running different hardware with only the OS being common between the two... I just don't believe in coincidences.

I'll report back on what SMART finds, though.

[TheFu]

Samsung 850 EVO had some issues a few years ago. A firmware update to the drive or bios was needed. Sorry, I don't recall which.

Also, check the system log files for errors and warnings. That could confirm the action in the kernel causing problems, RAM or some other issue. A few weeks ago, I had to slow the RAM in my Ryzen box from 2800 Mhz to 2733 Mhz to prevent an page swap kernel trace/crash. About a year ago, I had to reduce the RAM speed from 2966 to 2800 just to get a stable system after doubling the RAM.

[rebeltaz]

Samsung 850 EVO had some issues a few years ago. A firmware update to the drive or bios was needed. Sorry, I don't recall which.

Also, check the system log files for errors and warnings. That could confirm the action in the kernel causing problems, RAM or some other issue. A few weeks ago, I had to slow the RAM in my Ryzen box from 2800 Mhz to 2733 Mhz to prevent an page swap kernel trace/crash. About a year ago, I had to reduce the RAM speed from 2966 to 2800 just to get a stable system after doubling the RAM.

I can look at the log files (actually I just took a look) but I have NO clue what I am looking at! I looked at syslog.1 (because that one was from yesterday when it actually locked up) and the only errors I see have to do with a video downloader I run.

[TheFu]

syslog is just 1 log file. There are others. Check out journalctl to see the hardware logs. That command has all sorts of filter capabilities. Time, subsystem, errors are just some of the filters. Mine is setup to retain the last 10 boots of logs. I don't know what the default is.

[Topsiho]

TheFu is probably much more experienced than I am. Still I think it strange that the same problems exist on two different computers. Myself am always thinking of hardware problems rather than the Ubuntu software, but not this time.
I found crash reports in /var/crash, all very recent and with the same kernel. But also all different. On my wife's laptop (thus the third system here) with problems since probably a week. And none for years before that. On my other systems, with ubuntu 20.04 /var/crash is empty.

So it is almost certain the kernel is at fault, this time. No need to test the hardware: disks, memory and such, in this special situation.
My wife now boots from another kernel and so far reports no problems.

So I think in this very special case pointing a finger to the kernel is not so strange.

Topsiho

[dragonfly41]

Either point at a common factor such as kernel
or
if malicious files are passed between these two devices
it might pay to check both devices as documented here.

sudo rkhunter --checkall

sudo cat /var/log/rkhunter.log | grep -i warning

and possibly ... for deeper test ...

sudo apt-get install chkrootkit

[rebeltaz]

There seems to be another kernel available for upgrade (153) so... that might be a sign of issues with the previous one?

[TheFu]

There seems to be another kernel available for upgrade (153) so... that might be a sign of issues with the previous one?

All software has issues. It is highly unlikely that any specific fix will help you. The changes for every new kernel are spelled out in the kernel changelog. This is public information. Each week, a post of all Ubuntu updates is made to these forums, if you don't want to read from the actual changelog. I usually skim that list. I patch weekly, so it is clearer when an issue is related to patching or something else. If my system starts behaving funny on a Tuesday, then it isn't from patching. I patch on Saturday mornings.

[rebeltaz]

All software has issues. It is highly unlikely that any specific fix will help you. The changes for every new kernel are spelled out in the kernel changelog. This is public information. Each week, a post of all Ubuntu updates is made to these forums, if you don't want to read from the actual changelog. I usually skim that list. I patch weekly, so it is clearer when an issue is related to patching or something else. If my system starts behaving funny on a Tuesday, then it isn't from patching. I patch on Saturday mornings.

Be that as it may... all I can say is that since upgrading to a new kernel, neither computer has had a single issue since! I appreciate all of the help.

[TheFu]

Be that as it may... all I can say is that since upgrading to a new kernel, neither computer has had a single issue since! I appreciate all of the help.

Being lucky is a good thing! Happy you found the problem and the solution was available. I have a few kernel issues which are known and unsolved since 2016.