Once upon a time, Robert <kerplop@xxxxxxxxxxxxx> said:
After consulting man tune2fs, I follow what you're doing above. I remain mystified about the source of your information, though.
When I ran FC2 parted to resize a partition a week or two ago, it said that the problem was the dir_index option, and I happened to still remember it when I saw your email. :-)
Thanks again for remembering but I have yet the experience the "Thrill of victory". I tried the sequence you spelled out without success and decided it just might be the version of parted and/or tune2fs available in FC1. For the record, that's parted 1.6.3-31 and tune2fs 1.34 (25-Jul-2003). So, I got out my trusty FC2 rescue CD and booted this machine from it. (Skipped network setup, skipped mounting / fs on /mnt/sysimage.) This gives me tune2fs 1.35 (28 Feb 2004) and GNU parted 1.6.9 to work with and I got the same results. First I tried tune2fs \ -O ^dir_index /dev/hdb2 then tried the same command on the /boot part (hdb1) too. Parted complained, either way, with the same msg as before.
I'm thinking that the tune2fs command isn't doing anything because when I theoretically restored the dir_index for the two partitions, any delay before the bash prompt returned was imperceptable.
Y'know, since removing the rescue CD and rebooting to FC1, I thought of something else that's not exactly normal. When I loaded FC2 onto that drive, I unplugged the pri master drive to be damn certain that any misunderstanding 'tween Anaconda and I would cost nothing but time. Therefore GRUB was installed on the pri slave drive.
I wonder if that's spoiling my party... That's for tomorrow.
-- HOST SYSTEM RESPONDING, PROBABLY UP...
20:24:00 up 4 min, 2 users, load average: 0.47, 0.42, 0.18
One billion seconds ago it was 17:37:20 CST Mon 11/13/72