e2fsck, LVM and 4096-char block problems

e2fsck, LVM and 4096-char block problems

[Mike Castle]

Using Linus' 2.4.10, unpatched.  (Perhaps I need to patch the LVM stuff ;-)

I wanted to extend /usr/src from 19G to 20G.

I umounted /usr/src, lvextend -L +1G /dev/vg01/src, fsck -f /usr/src,
resize2fs /dev/vg01/src

resize2fs tells me to run fsck.  Hmmmm... I just did.

I poke around a bit more and notice the following:

Oct 19 11:45:41 thune kernel: ll_rw_block: device 3a:00: only 4096-char
blocks implemented (1024)

Of course, dumpe2fs says:
Block size:               4096

I tried -f on the resize, and got LOTS of the 4096-char messages.

Some perhaps pertinent information.

thune:~# vgdisplay /dev/vg01
--- Volume group ---
VG Name               vg01
VG Access             read/write
VG Status             available/resizable
VG #                  0
MAX LV                256
Cur LV                9
Open LV               8
MAX LV Size           255.99 GB
Max PV                256
Cur PV                5
Act PV                5
VG Size               150.29 GB
PE Size               4 MB
Total PE              38473
Alloc PE / Size       28746 / 112.29 GB
Free  PE / Size       9727 / 38 GB
VG UUID               xdzl14-LsZ2-IA1c-dGk4-F8EK-xPlt-DV31mQ


thune:~# lvdisplay /dev/vg01/src
--- Logical volume ---
LV Name                /dev/vg01/src
VG Name                vg01
LV Write Access        read/write
LV Status              available
LV #                   1
# open                 0
LV Size                19 GB
Current LE             4864
Allocated LE           4864
Allocation             next free
Read ahead sectors     120
Block device           58:0


Obviously I have the option of creating a new volume and moving things over
to that, but that sort of defeats the purpose of these particular tools.

Recommended course of action?

mrc
-- 
     Mike Castle      dalgoda@ix.netcom.com      www.netcom.com/~dalgoda/
    We are all of us living in the shadow of Manhattan.  -- Watchmen
fatal ("You are in a maze of twisty compiler features, all different"); -- gcc

Re: e2fsck, LVM and 4096-char block problems

[Andreas Dilger]

On Oct 19, 2001  11:59 -0700, Mike Castle wrote:
> Using Linus' 2.4.10, unpatched.  (Perhaps I need to patch the LVM stuff ;-)

Very bad combination.  Don't use 2.4.10, don't use stock Linus LVM.

> resize2fs tells me to run fsck.  Hmmmm... I just did.

Fsck may not work on 2.4.10.

> Oct 19 11:45:41 thune kernel: ll_rw_block: device 3a:00: only 4096-char
> blocks implemented (1024)

Known problem fixed in more recent LVM.

Cheers, Andreas
--
Andreas Dilger  \ "If a man ate a pound of pasta and a pound of antipasto,
                 \  would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/               -- Dogbert

Re: e2fsck, LVM and 4096-char block problems

[Mike Castle]

On Fri, Oct 19, 2001 at 01:25:13PM -0600, Andreas Dilger wrote:
> On Oct 19, 2001  11:59 -0700, Mike Castle wrote:
> > Using Linus' 2.4.10, unpatched.  (Perhaps I need to patch the LVM stuff ;-)
> 
> Very bad combination.  Don't use 2.4.10, don't use stock Linus LVM.

I hadn't built 2.4.12 yet.

What needs to be done to get recent LVM into stock kernel?

I feel like it's the 2.2 RAID stuff all over again.

mrc
-- 
     Mike Castle      dalgoda@ix.netcom.com      www.netcom.com/~dalgoda/
    We are all of us living in the shadow of Manhattan.  -- Watchmen
fatal ("You are in a maze of twisty compiler features, all different"); -- gcc

Re: e2fsck, LVM and 4096-char block problems

[Andreas Dilger]

On Oct 19, 2001  12:35 -0700, Mike Castle wrote:
> > On Oct 19, 2001  11:59 -0700, Mike Castle wrote:
> > > Using Linus' 2.4.10, unpatched.  (Perhaps I need to patch the LVM stuff ;-)
> > Very bad combination.  Don't use 2.4.10, don't use stock Linus LVM.
> 
> What needs to be done to get recent LVM into stock kernel?
> 
> I feel like it's the 2.2 RAID stuff all over again.

Well, it is mostly up-to-date in -ac, but I think the Sistina folks have
given up on submitting it to Linus.  I do a bit now and then, but haven't
made much of a real effort to get it fully up-to-date.  Maybe now that the
whole block-device changes have settled down again I will submit a few
patches.

Cheers, Andreas
--
Andreas Dilger  \ "If a man ate a pound of pasta and a pound of antipasto,
                 \  would they cancel out, leaving him still hungry?"
http://www-mddsp.enel.ucalgary.ca/People/adilger/               -- Dogbert