[linux-lvm] multiple lvm2-snapshots on the same origin -> device-mapper error

[linux-lvm] multiple lvm2-snapshots on the same origin -> device-mapper error

[Patrick Reinhardt]

Hi!
I'm running in trouble creating more than 16 snapshots of the same origin:

--

# /sbin/lvcreate  -s -L150M -n rootfs_node17 /dev/vg/ro_rootfs
File descriptor 5 left open
File descriptor 6 left open
  Rounding up size to full physical extent 152.00 MB
  device-mapper ioctl cmd 9 failed: Cannot allocate memory
  Couldn't load device 'vg-rootfs_node8'.
  Problem reactivating origin ro_rootfs

# lvm version
  LVM version:     2.01.04 (2005-02-09)
  Library version: 1.01.00-ioctl (2005-01-17)
  Driver version:  4.4.0

# uname -a
  Linux hostname 2.6.11.12-xen0 #1 Thu Aug 4 00:45:12 BST 2005 i686
GNU/Linux

--

Browsing the mailing list archives I didn't find a suitable solution
although the problem was mentioned before, e.g. on
http://lists.xensource.com/archives/html/xen-devel/2004-09/msg00426.html

Any ideas?

Thank you in advance.

Patrick Reinhardt

Re: [linux-lvm] multiple lvm2-snapshots on the same origin -> device-mapper error

[Alasdair G Kergon]

On Tue, Dec 13, 2005 at 12:47:07PM +0100, Patrick Reinhardt wrote:
> Hi!
> I'm running in trouble creating more than 16 snapshots of the same origin:
 
We don't have a solution for that particular problem yet, but we do have code 
changes that should avoid the risk of the system locking up when it happens.

This needs the latest development packages for LVM2 (2.02.*) and 
device-mapper (1.02.*).

And a kernel containing the patches:

  dm-rename-frozen_bdev.patch
  dm-skip-lockfs.patch
  dm-ioctl-skip-lockfs.patch
  dm-snapshot-metadata-reading-separation.patch
  dm-snapshot-load-metadata-on-creation.patch
  dm-ioctl-reduce-pf-memalloc-usage.patch

from http://www.kernel.org/pub/linux/kernel/people/agk/patches/2.6/editing/

There are still some snapshot problems that have shown up while
testing the RHEL4 kernel (with all the above patches applied) that we 
haven't got to the bottom of yet:

  https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=174636
  https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=174742
  https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=174753

Alasdair
-- 
agk@redhat.com

Re: [linux-lvm] multiple lvm2-snapshots on the same origin -> device-mapper error

[Patrick Reinhardt]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi!
Thank you for the hints.

Alasdair G Kergon schrieb:
> We don't have a solution for that particular problem yet, but we do have code 
> changes that should avoid the risk of the system locking up when it happens.

In fact i need about 40 snapshots of a single filesystem for xen. Do you
have any idea how to circumvent the problem? Maybe by creating 3
identical origin filesystems and distribute the snapshots?

Thank you in advance.

Patrick Reinhardt
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3-cvs (MingW32)
Comment: GnuPT 2.7.4

iD8DBQFDns/rf1C3plDEOPkRAuUpAJsHKnUEY+ZqQAU7f5WVBkK1lWZWfwCePLbQ
iF2UzozJ4yLUM4ZUoTVwlg4=
=i8JQ
-----END PGP SIGNATURE-----

Re: [linux-lvm] multiple lvm2-snapshots on the same origin -> device-mapper error

[Patrick Reinhardt]

Hi!
Thank you for the hints.

Alasdair G Kergon schrieb:

> We don't have a solution for that particular problem yet, but we do
have code
>changes that should avoid the risk of the system locking up when it
happens.


In fact I need about 40 snapshots of a single filesystem for xen. Do you
have any idea how to circumvent the problem? Maybe by creating 3
identical origin filesystems and distribute the snapshots?

Thank you in advance.

Patrick Reinhardt

Re: [linux-lvm] multiple lvm2-snapshots on the same origin -> device-mapper error

[Klaus Strebel]

Patrick Reinhardt schrieb:
> Hi!
> Thank you for the hints.
> 
> Alasdair G Kergon schrieb:
> 
> 
>>We don't have a solution for that particular problem yet, but we do
> 
> have code
> 
>>changes that should avoid the risk of the system locking up when it
> 
> happens.
> 
> 
> In fact I need about 40 snapshots of a single filesystem for xen. Do you
> have any idea how to circumvent the problem? Maybe by creating 3
> identical origin filesystems and distribute the snapshots?
Looks to me like you really want a kind of cluster filesystem ...

Ciao
Klaus
-- 
Mit freundlichen Gr�ssen / best regards

Klaus Strebel, Dipl.-Inform. (FH), mailto:klaus.strebel@gmx.net

/"\
\ /     ASCII RIBBON CAMPAIGN
  X        AGAINST HTML MAIL
/ \