Shared publicly  - 
 
I have never seen something like this before.

This file is not in use by anything else on the system.

$ md5sum xppro.vdi
589cbb5501dcddda047344a3550aaa95 xppro.vdi
$ md5sum xppro.vdi
a69806ec60d39e06473edbb0abd71637 xppro.vdi

No errors on dmesg, no explanation for this. The behavior is repeatable - every invocation produces a different result. It behaves the same regardless of which partition I put the file on.

More details on my blog at http://changelog.complete.org/archives/7452-unreported-disk-data-corruption-kernel-bug . Any ideas for troubleshooting welcome, of course.
2
Eric Gillespie's profile photoDan Piponi's profile photoAndrew Moss's profile photoYesudeep Mangalapilly's profile photo
12 comments
 
I've seen varying md5s for the same file with a USB flash drive that had been through the washing machine.
 
The only time I've ever seen that result is when one stick of memory had errors with only one particular pattern.
I'd strongly suggest you run memtest86+ overnight on that system.
 
Excellent idea, +Shae Erisson. I don't know why I didn't think of that. It detected errors within 30 seconds of invocation. I rebooted it a few minutes later and it detected errors at the exact same spot in RAM again that time. Smoking gun found.
 
Now I'm debating whether to nuke my Debian install and restart from scratch. That's a lot of time down the drain but might be worth it. I have no idea what could have been corrupted. debsums could verify my installed files, of course. I don't have much in /home that would go unnoticed, or so I hope. and maybe some fsck's would catch any fs issues. What do you think?
 
I had exactly that same problem, with the md5sum result being different on every run, so I've seen this before.
I guess hash functions are an effective way to find memory errors at runtime :-)
 
You could have errors in generated files such as the symlinks for the preferences system (vi runs which installed binary?), those wouldn't be caught by debsums. I'd nuke it.
 
+John Goerzen can you just tell it to re-install everything rather than having to do all the formatting + configuring again?
 
Once I had a situation where I kept getting random kernel panics. I changed the memory, reinstalled Linux, but only when I used another machine did I stumble upon a faulty spike suppressor. Also, realized that if you have a square wave inverter as your power backup, you'll likely face such issues as well. I don't know if this helps but I'd definitely replace RAM in this case.
 
You should reformat as well as replacing all of the files. Which fs are you using? The ext family uses superblock markers scattered throughout the disk. You really don't want these to have been corrupted when the disk was formatted last time...
 
That is the plan. Repartition, reformat, etc. I had the installer resize the Windows partition too, so basically everything on disk is suspect. Going to completely redo it all.
 
Nope. Even if it was, it should md5sum the same each time though.
Add a comment...