[ext2resize] Load testing: ext2resize complains about a bug

[Poul P. <pe...@ro...>]

	I've been running some tests with online resizing of ext3 volumes.
First, let me explain the setup:

kernel 2.4.19
lvm-1.0.6
ext3-2.4 + online_resize patch
ext2resize from CVS (~12/6)
RedHat 7.2 running on dual P-II 400
Qlogic 2200 San -> Zzyzx RocketStor 2000 

	Through the san I assigned one 205GB raid set to this machine and
setup a 1GB LV and ext3 filesystem. I then mounted this filesystem, exported
it via NFS, and set about 16 hosts to read and write to the volume (we have
a home-brewed nfsload script that tries to mimic the build activity we see
on out production NFS server). I then set e2fsadm in a loop adding 512MB to
the volume every 5 minutes. The server never crashed or experienced any
difficulties and the LV was eventually extended to the full size of the VG.
However, the ext3 file-system stopped extending at about 16GB. I stopped the
activity, unmounted the disk and ran e2fsck which reported no errors. I then
tried to do a ext2resize with the disk offline and got the following:

[root@nfs1node1 ext2resize-1.1.18_pre]# ext2resize /dev/vg00/sa_test
ext2resize v1.1.18 - 2001/03/18 for EXT2FS 0.5b
BUG! block 32782 shouldn't have been marked!
Please send a filesystem dump and ext2resize version to:
<ext...@li...>

	Naturally, I'm not about to send a 16GB dump to the mailing list :)
I have not done anything else to this filesystem and the hardware is all
test hardware, so I can do pretty much anything that would be helpful.
Thanks for any suggestions...

-poul