[linux-lvm] LVM and resize

[linux-lvm] LVM and resize

[Reidar Westvik]

[-- Attachment #1: Type: text/plain, Size: 1753 bytes --]

Hi,

I don't know if my question in something that Sistina supports, but i need some pointers to make this work.

We're running a Slackware 9.0 server with kernel 2.5.64 (gcc ver. 2.95.3 20010315) also on this server we have a LVM volume of 1TB spread over 120GB IDE disks. Now, it seems like one of the harddrives has failed in some way. Filesystem on the volume is EXT3.

When i mount the volume syslog says: "kernel: EXT3-fs warning: mounting fs with errors, running e3fschk is recommended" So i've unmounted the volume and run e2fschk on the LVM volume. After a long wait (2 days in fact) the server hangs, when its checking the part of the volume that is on hdf. The box just dies, and only a powercycle will bring it back up again. The syslogs displays the following info:
"kernel: hdf: dma_timer_expiry: dma status == 0x61"
"kernel: hdf: (__ide_dma_test_irq) called while not waiting""

These are the last events of syslog before it's starts to reboot.

Now, after what i've read the way to deal with this issue without loss of data is to shrink the volume with aprox. 120gb (there is enough free space) and then exclude the defective drive from the logical volume and its physical groups, and then add a healthy disk for so to increase the volume back to its original size.

However LVM resize tools will not allow resizing of a volume that has a filsystem marked with errors. And if i try to force the resize the server hangs. I have tryed to resize with both "e2fsadm" and "resize2fs" but they both have problem resizing since it has errors.

So this kinda puts me back to basics, is there a way for me to remove this disk without loosing data on the volume ?

All awnsers are very much apriciated! :)

Best Regards
Reidar

[-- Attachment #2: Type: text/html, Size: 3059 bytes --]

Re: [linux-lvm] LVM and resize

[Heinz J . Mauelshagen]

Looks like you are caught IRT to shrinking, because of the fsck errors.

If you're still able to mount the filesystem, you definitely want to take a backup
first in order to preserve as much of your files as possible.

Aftwards (presumably you can add another >=120g drive to your VG with vgextend) you
want to try moving the data off to that other disk using "pvmove -i ...", then
removing the failing one (vgreduce).

Using device-mapper/LVM2, which enables you to remove the drive in question
(removing LVs allocated to that drive as well), doesn't help you here, presumably
you just have that one huge Logical Volume.

Regards,
Heinz    -- The LVM Guy --

On Mon, Oct 06, 2003 at 04:25:39PM +0200, Reidar Westvik wrote:
> Hi,
> 
> I don't know if my question in something that Sistina supports, but i need some pointers to make this work.
> 
> We're running a Slackware 9.0 server with kernel 2.5.64 (gcc ver. 2.95.3 20010315) also on this server we have a LVM volume of 1TB spread over 120GB IDE disks. Now, it seems like one of the harddrives has failed in some way. Filesystem on the volume is EXT3.
> 
> When i mount the volume syslog says: "kernel: EXT3-fs warning: mounting fs with errors, running e3fschk is recommended" So i've unmounted the volume and run e2fschk on the LVM volume. After a long wait (2 days in fact) the server hangs, when its checking the part of the volume that is on hdf. The box just dies, and only a powercycle will bring it back up again. The syslogs displays the following info:
> "kernel: hdf: dma_timer_expiry: dma status == 0x61"
> "kernel: hdf: (__ide_dma_test_irq) called while not waiting""
> 
> These are the last events of syslog before it's starts to reboot.
> 
> Now, after what i've read the way to deal with this issue without loss of data is to shrink the volume with aprox. 120gb (there is enough free space) and then exclude the defective drive from the logical volume and its physical groups, and then add a healthy disk for so to increase the volume back to its original size.
> 
> However LVM resize tools will not allow resizing of a volume that has a filsystem marked with errors. And if i try to force the resize the server hangs. I have tryed to resize with both "e2fsadm" and "resize2fs" but they both have problem resizing since it has errors.
> 
> So this kinda puts me back to basics, is there a way for me to remove this disk without loosing data on the volume ?
> 
> All awnsers are very much apriciated! :)
> 
> Best Regards
> Reidar

*** Software bugs are stupid.
    Nevertheless it needs not so stupid people to solve them ***

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Re: [linux-lvm] LVM and resize

[Reidar Westvik]

Hi Heinz,

Thanks for your quick reply! :)

I am still able to mount the filesystem, but after a while it goes into
read-only mode, witch is fine. However I cannot take backup of the volume
cause it's to big, but it's not "the end of the world" if the data gets
lost, but we would like to keep data.

I've read about using pvmove -i and when i try to run it i get "command not
implemented yet". Is there a way around this ?

Reidar
----- Original Message ----- 
From: "Heinz J . Mauelshagen" <mauelshagen@sistina.com>
To: <linux-lvm@sistina.com>
Sent: Tuesday, October 07, 2003 3:53 PM
Subject: Re: [linux-lvm] LVM and resize


>
> Looks like you are caught IRT to shrinking, because of the fsck errors.
>
> If you're still able to mount the filesystem, you definitely want to take
a backup
> first in order to preserve as much of your files as possible.
>
> Aftwards (presumably you can add another >=120g drive to your VG with
vgextend) you
> want to try moving the data off to that other disk using "pvmove -i ...",
then
> removing the failing one (vgreduce).
>
> Using device-mapper/LVM2, which enables you to remove the drive in
question
> (removing LVs allocated to that drive as well), doesn't help you here,
presumably
> you just have that one huge Logical Volume.
>
> Regards,
> Heinz    -- The LVM Guy --
>
> On Mon, Oct 06, 2003 at 04:25:39PM +0200, Reidar Westvik wrote:
> > Hi,
> >
> > I don't know if my question in something that Sistina supports, but i
need some pointers to make this work.
> >
> > We're running a Slackware 9.0 server with kernel 2.5.64 (gcc ver. 2.95.3
20010315) also on this server we have a LVM volume of 1TB spread over 120GB
IDE disks. Now, it seems like one of the harddrives has failed in some way.
Filesystem on the volume is EXT3.
> >
> > When i mount the volume syslog says: "kernel: EXT3-fs warning: mounting
fs with errors, running e3fschk is recommended" So i've unmounted the volume
and run e2fschk on the LVM volume. After a long wait (2 days in fact) the
server hangs, when its checking the part of the volume that is on hdf. The
box just dies, and only a powercycle will bring it back up again. The
syslogs displays the following info:
> > "kernel: hdf: dma_timer_expiry: dma status == 0x61"
> > "kernel: hdf: (__ide_dma_test_irq) called while not waiting""
> >
> > These are the last events of syslog before it's starts to reboot.
> >
> > Now, after what i've read the way to deal with this issue without loss
of data is to shrink the volume with aprox. 120gb (there is enough free
space) and then exclude the defective drive from the logical volume and its
physical groups, and then add a healthy disk for so to increase the volume
back to its original size.
> >
> > However LVM resize tools will not allow resizing of a volume that has a
filsystem marked with errors. And if i try to force the resize the server
hangs. I have tryed to resize with both "e2fsadm" and "resize2fs" but they
both have problem resizing since it has errors.
> >
> > So this kinda puts me back to basics, is there a way for me to remove
this disk without loosing data on the volume ?
> >
> > All awnsers are very much apriciated! :)
> >
> > Best Regards
> > Reidar
>
> *** Software bugs are stupid.
>     Nevertheless it needs not so stupid people to solve them ***
>
>
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
=-
>
> Heinz Mauelshagen                                 Sistina Software Inc.
> Senior Consultant/Developer                       Am Sonnenhang 11
>                                                   56242 Marienrachdorf
>                                                   Germany
> Mauelshagen@Sistina.com                           +49 2626 141200
>                                                        FAX 924446
>
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
=-
>
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
>
>
>

Re: [linux-lvm] LVM and resize

[Heinz J . Mauelshagen]

On Tue, Oct 07, 2003 at 07:13:14PM +0200, Reidar Westvik wrote:
> Hi Heinz,
> 
> Thanks for your quick reply! :)
> 
> I am still able to mount the filesystem, but after a while it goes into
> read-only mode, witch is fine. However I cannot take backup of the volume
> cause it's to big, but it's not "the end of the world" if the data gets
> lost, but we would like to keep data.
> 
> I've read about using pvmove -i and when i try to run it i get "command not
> implemented yet". Is there a way around this ?

You mean with LVM2 ? Not yet.
Will probably be in 2.2.

Regards,
Heinz    -- The LVM Guy --


> 
> Reidar
> ----- Original Message ----- 
> From: "Heinz J . Mauelshagen" <mauelshagen@sistina.com>
> To: <linux-lvm@sistina.com>
> Sent: Tuesday, October 07, 2003 3:53 PM
> Subject: Re: [linux-lvm] LVM and resize
> 
> 
> >
> > Looks like you are caught IRT to shrinking, because of the fsck errors.
> >
> > If you're still able to mount the filesystem, you definitely want to take
> a backup
> > first in order to preserve as much of your files as possible.
> >
> > Aftwards (presumably you can add another >=120g drive to your VG with
> vgextend) you
> > want to try moving the data off to that other disk using "pvmove -i ...",
> then
> > removing the failing one (vgreduce).
> >
> > Using device-mapper/LVM2, which enables you to remove the drive in
> question
> > (removing LVs allocated to that drive as well), doesn't help you here,
> presumably
> > you just have that one huge Logical Volume.
> >
> > Regards,
> > Heinz    -- The LVM Guy --
> >
> > On Mon, Oct 06, 2003 at 04:25:39PM +0200, Reidar Westvik wrote:
> > > Hi,
> > >
> > > I don't know if my question in something that Sistina supports, but i
> need some pointers to make this work.
> > >
> > > We're running a Slackware 9.0 server with kernel 2.5.64 (gcc ver. 2.95.3
> 20010315) also on this server we have a LVM volume of 1TB spread over 120GB
> IDE disks. Now, it seems like one of the harddrives has failed in some way.
> Filesystem on the volume is EXT3.
> > >
> > > When i mount the volume syslog says: "kernel: EXT3-fs warning: mounting
> fs with errors, running e3fschk is recommended" So i've unmounted the volume
> and run e2fschk on the LVM volume. After a long wait (2 days in fact) the
> server hangs, when its checking the part of the volume that is on hdf. The
> box just dies, and only a powercycle will bring it back up again. The
> syslogs displays the following info:
> > > "kernel: hdf: dma_timer_expiry: dma status == 0x61"
> > > "kernel: hdf: (__ide_dma_test_irq) called while not waiting""
> > >
> > > These are the last events of syslog before it's starts to reboot.
> > >
> > > Now, after what i've read the way to deal with this issue without loss
> of data is to shrink the volume with aprox. 120gb (there is enough free
> space) and then exclude the defective drive from the logical volume and its
> physical groups, and then add a healthy disk for so to increase the volume
> back to its original size.
> > >
> > > However LVM resize tools will not allow resizing of a volume that has a
> filsystem marked with errors. And if i try to force the resize the server
> hangs. I have tryed to resize with both "e2fsadm" and "resize2fs" but they
> both have problem resizing since it has errors.
> > >
> > > So this kinda puts me back to basics, is there a way for me to remove
> this disk without loosing data on the volume ?
> > >
> > > All awnsers are very much apriciated! :)
> > >
> > > Best Regards
> > > Reidar
> > _______________________________________________
> > linux-lvm mailing list
> > linux-lvm@sistina.com
> > http://lists.sistina.com/mailman/listinfo/linux-lvm
> > read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
> >
> >
> >
> 
> 
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Sistina Software Inc.
Senior Consultant/Developer                       Am Sonnenhang 11
                                                  56242 Marienrachdorf
                                                  Germany
Mauelshagen@Sistina.com                           +49 2626 141200
                                                       FAX 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-