On Wed, Mar 17, 2010 at 02:39:58PM -0400, ty...@mit.edu wrote: > On Wed, Mar 17, 2010 at 12:16:01PM +0100, Wouter Verhelst wrote: > > Package: e2fsprogs > > Version: 1.41.10-1 > > Severity: normal > > > > A few days back, I tried to run 'resize2fs -M' on an ext4 filesystem. > > While it was running, I accidentally hit ^C in its xterm window (I was > > confused about which window had the focus). As a result, the resize was > > aborted. > > > > The filesystem contains a low number of DV streams (from a dvswitch > > recording). When I mounted it after this interrupted resize operation, > > one of the files in this filesystem was corrupt, and the kernel mentions > > that the filesystem has errors. I have only mounted it once or twice > > since the corruption occurred, and have not written anything to the > > filesystem; since then, I only used LVM snapshots. > > > > An 'e2fsck' or an 'e2fsck -f' on an LVM snapshot of the filesystem does > > not fix the problems. > > > > If you need, I can upload an e2image -r file that I have prepared > > somewhere. It's 101M after gzip, though. > > A URL to the e2image -r file (sent to me privately if you have privacy > concerns) would be appreciated, thanks.
http://flamenco.grep.be/~wouter/baobab-image.gz (still uploading at the time of writing, with an ETA of about 40 minutes) > I'm not sure there's a lot I'll be able to find from the e2image file, > however. > > What would be more important is knowing what resize2fs was doing (the > -p option would have been really helpful). I'm afraid I cannot tell you that anymore; I only found out about the corruption after the window in which the resize2fs had been running, was closed. > Also, using resize2fs -M has always been a bit dangerous with respect > to being interrupted. I'm not sure we'll be able to do much other > than put warnings in the man page. It might perhaps be good, then, to put a short warning in the output of resize2fs too, followed by a timeout to allow the user to interrupt while it is still safe. > If you could do some work to try to create a replicable test case > (standard file system image created using e2image -r as a 'before' > snapshot, instructions on when to interrupt resize2fs -pM, etc.) that > would be really helpful. I could try that, but I'm not sure it will be very successful. If you could first give the image that I've put online a look, I'll happily give this a shot if that doesn't help you. Thanks, -- The biometric identification system at the gates of the CIA headquarters works because there's a guard with a large gun making sure no one is trying to fool the system. http://www.schneier.com/blog/archives/2009/01/biometrics.html -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
