Al Sparks spake the following on 5/3/2007 6:39 PM:
--- Matt Hyclak hyclak@math.ohiou.edu wrote:
On Wed, May 02, 2007 at 11:06:31PM -0700, Al Sparks enlightened us:
Did you resize the filesystem, too?
Matt
Nope. How do you do that?
resize2fs would be a good guess. Usually this is done *before* you shrink the disk out from underneath it. If anything was in those sectors you removed from the LV, you might be out of luck.
Matt
Well, I was able to get my server back up.
First, I had to boot it up on a recovery CD (standard CentOS) and comment out the bad volume it wanted to mount in /etc/fstab.
When the server came back up, I tried using resize2fs to resize. The resize2fs would not let me resize until I manually ran fsck because the file system and logical volume block counts wouldn't match (been there, done that, no joy). So I increased the logical volume back to where it was to make them match, ran fsck again, and everything checked out. I then was able to use resize2fs to decrease the file system size, and then run lvreduce to decrease the volume size.
However, because neither application didn't seem to allow me to use block size parameters, I used units of Megabytes and would end up within a few blocks of each other. I then would estimate how much more units of +/- MB I needed to get things in sync.
Kind of a pain. Anyway, I appreciate the help. Nice to have these tools available to use. === Al
If it worked for you then you could resize2fs to "smaller" than you need, resize the LV, and resize2fs with no size parameters I believe will resize to all available space in the filesystem. Yes, it is one extra step, but it should work. I haven't used resize2fs for so long, I did't think it worked on ext3. I guess past pains don't always carry to the future.