Comment 149 for bug 509180

As for the bug, I forgot to mention that my disk may well have been filled 100% right before the error occurred. I think somebody posting above reported a similar thing. So maybe that can explain why the supposedly fixed bug resurfaced on my system using the 3.0.0-12 kernel?

For the record, to clean up the corrupted files, I used a variant of the tip somebody above posted to find and remove corrupted files:

find . -type f -size -10k -exec cat {} \; > /dev/null

To explain for dummies like me: This reads all files smaller than 10k and yields an Input/Output error for the files concerned. (which in my case were all 0-bytes files, mostly lockfiles). You can then go about removing those files manually.