Default ext3 Error Behavior

Latest response

I'd like some insight as to what the implications of setting the ext3 error-behavior flag to continue.  I believe by default this option is set to remount read-only, what is this really protecting against?


My application requires keeping a single OS up if at all possible, and minor log data (the only thing being written to the disk) corruption is acceptable.  What I do not want is a single write error causing the filesystem to be remounted read only.




AFAIK, the 'continue' error behaviour is broken upstream and so it works in the same way as 'remount-ro'.


By default, the behaviour is 'Continue' and since this is broken, it falls back to 'remount-ro'.


On a test machine :


# tune2fs -l /dev/sda1 | grep -i continue
Errors behavior:          Continue