Because I still haven’t remembered to set myself up with a limited-access account for blogging on my new site, I’m back here again. (EDIT: Migrated from old site)

This time is a lot of fun and games with removing an “open” logical volume from within a CentOS environment. The dom0 machine has a set of logical volumes in a volume group for each of the VMs and one of them went and failed dismally, and after its xen destruction (xm destory foo) its LV was left with it being “open”.

It’s not exactly easy to see how/why this would be the case and every command (lsof, fuser, mount, etc) told me that the LV wasn’t in use. But still, lvremove, lvchange, and dmsetup commands failed.

But, I’ve solved my problem! Hopefully yours too.. read on through the commands:

# lvchange -an /dev/vmsvg1/foo
LV vmsvg1/foo in use: not deactivating
# dmsetup info -c vmsvg1-foo
Name Maj Min Stat Open Targ Event UUID
vmsvg1-foo 253 65 L–w 1 1 2 XXX
# dmsetup remove -force vmsvg1-foo
device-mapper: remove ioctl failed: Device or resource busy
# lvchange -an /dev/vmsvg1/foo
/dev/vmsvg1/foo: read failed after 0 of 4096 at 0: Input/output error
/dev/mapper/vmsvg1-foop1: read failed after 0 of 2048 at 0: Input/output error
# dmsetup remove -force vmsvg1-foop1
# dmsetup remove -force vmsvg1-foo
# lvs vmsvg1
foo vmsvg1 -wi— 4.00G (it’s closed!!)
# lvchange -an /dev/vmsvg1/foo
# lvremove /dev/vmsvg1/foo
Logical volume “foo” successfully removed

So, after all that, it looks like there’s some stupid ‘p1′ references in /dev/mapper/ . Nfi why, but removing it (albeit by force with dmsetup) solved the problem.

So it’s not that special and I bet 1000 people have already posted the same details. Still, it’s cool and I need a place to record my thoughts about this.

ssh host.name -L YYYY:other.host:ZZZZ

Essentially, this means logging into the first machine and creating port YYYY on the local machine as the port ZZZZ from other.host

Works well.