Bug#421550: Issues at shutdown when using a mix of local and network accessible block devices.

[jacques Normand]

Package: lvm2
Architecture: amd64
Version: 2.02.06-4

Severity: grave


Hi,

I am having an issue with lvm2 at shutdown. I am using an lvm over raid1
setup, all on local sata drives. There is no problems as long as I do
not try to use my external storage (coraid bd with large volume group on
it). If I vgscan / vgchange -a y  that volume group, I will not be able
to shutdown properly anymore. That costed me a mangled /var partition,
which is why I placed a severity: grave. 

Now, what happens on the system:
* start the network
* load aoe module
* vgscan / vgchange
* mount the partition

* do some useful work

* umount the partition
* vgchange -a n 
* unload the aoe module (I had to add that to the init.d/aoetools, it
                         will be another bug report)
* shutdown the network
* watch LVM2 initscript lockup after displaying:
  Shutting down LVM Volume Groups... 

My guess is that is tries to shutdown the remote volumes even though
they are already disabled. If the aoe module where still loaded, I would
get a nice kernel panic (this is how I found that I had to unload the
module). But is there a way to remove a volume group from the list of
groups that lvm knows about, or better, get a list of active groups and
call vgchange -a n on each rather than the bulk, one call deal.

thanks

jacques

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.alioth.debian.org/pipermail/pkg-lvm-maintainers/attachments/20070429/aac530a9/attachment.pgp