I don't think this is related to #566329, but it might be. The reason I don't think it's related is that this occurred running on the 174.2.3 kernel, which seemingly isn't falling foul of #566329. This is on a fully up to date F12, building an F11 live CD: resize2fs 1.41.9 (22-Aug-2009) Resizing the filesystem on /var/tmp/imgcreate-47Szwr/tmp-s_adO_/ext3fs.img to 1667225 (4k) blocks. The filesystem on /var/tmp/imgcreate-47Szwr/tmp-s_adO_/ext3fs.img is now 1667225 blocks long. e2fsck 1.41.9 (22-Aug-2009) Pass 1: Checking inodes, blocks, and sizes Pass 2: Checking directory structure Pass 3: Checking directory connectivity Pass 4: Checking reference counts Pass 5: Checking group summary information Free inodes count wrong for group #32 (8192, counted=23). Fix? yes Free inodes count wrong (191598, counted=183429). Fix? yes JAH-live-F11-x86: ***** FILE SYSTEM WAS MODIFIED ***** JAH-live-F11-x86: 234363/417792 files (0.2% non-contiguous), 1656071/1667225 blocks /usr/lib/python2.6/site-packages/imgcreate/errors.py:45: DeprecationWarning: BaseException.message has been deprecated as of Python 2.6 return unicode(self.message) Error creating Live CD : fsck after resize returned an error! image to debug at /tmp/resize-image-pTYaoX I'll attach the debug image. James
Created attachment 395241 [details] screwed up resized image
Ah--I hadn't realised that a new build of e2fsprogs, and a new kernel, hit my machine overnight. So I was running e2fsprogs-1.41.9-6.fc12.x86_64, but without realising this. The bug report for #566329 would have been 1.41.9-5. I don't know if that's relevant. I hadn't rebooted, so I was still on the 174.2.3 kernel. I have rebooted the machine, still in the 174.2.3 kernel, just in case the old e2fs libs were already in memory. Running the build again, with the new e2fsprogs, but the old kernel. After that, I'll try the new e2fsprogs and the new kernel, and report back.
Failed on both attempts, I'm afraid. This is still happening with the new kernel and new e2fsprogs, so it is a genuine live bug, and different from #566329. Let me know if I can help in any way. James
OK, I think I have something more useful than the previous debug image. I'll attach another image, which gets screwed up if you try to resize it (using e2fsprogs-1.41.9-6.fc12.x86_64). The image as it stands checks fine. But if you then resize2fs -M bad-resize.img and then e2fsck again, you get errors. James
Created attachment 395259 [details] image that gets screwed by resize2fs -M
thanks, I guess ;) I'll look into this one. livecd really challenges resize2fs... -Eric
This problem affects me too... when trying to build a remix. I also added comments to #566329. Thanks.
if you narrow it down to either e2fsprogs, or kernel, please let me know; otherwise will try to get to this soon (but I'm actually trying to make my weekend a weekend ...)
It happens when invoking resize2fs, so surely that means it's e2fsprogs? It wasn't a great shock that #566329 was a kernel issue because it was a question of corruption on a mounted filesystem, i.e., it was the kernel filesystem driver's job to do the disk access and work out where to put things. But here, we're explicitly invoking resize2fs on an unmounted filesystem--so I would think the kernel wouldn't be involved. But I am a little out of my depth. Anyway, I'll reboot into 174.2.3 (which didn't manifest the error with the old e2fsprogs) and see if it now does (with the new e2fsprogs). (PS Weekends... I remember them...)
Can you try 1.41.9-7 from testing? I think 1.41.9-6 had a bug and I superceded it with -7: %changelog * Sun Feb 07 2010 Eric Sandeen <sandeen> 1.41.9-7 - Upstream version of resize2fs array fix, original was wrong I pushed -6 to stable instead of -7, arg...
I actually just pushed -7 to stable, but it'll take a little while to get out. Grr not ext4's best week!
Just finishing up the i686 remix build and it got past the problem part... so the -7 release of e2fsprogs seems to have taken care of the problem for me. I'll also try it on the x86_64 remix shortly. Thanks.
x86_64 worked great. Thanks.
> I actually just pushed -7 to stable, but it'll take a little while to get out. Perfick! Now building with no errors. Thank you. Will you also build for F11? > Grr not ext4's best week! It has finished the week the stronger for it. Good work (and sorry about using up your weekend). James
Ok, I'll mark fixed in e2fsprogs-1.41.9-7 I don't think this ever made it to F11 so probably no fixing there.