Red Hat Bugzilla – Bug 858110
change the device in the man page of thin_check, thin_dump & thin_restore
Last modified: 2015-01-26 19:10:53 EST
Description of problem:
thin_check manpage issue:
there is an example in thin_check.
it's better to change the device to /dev/mapper/vg-pool_tmeta since the internal thin meta device is not accessed in /dev/vg/.
same issues for thin_dump & thin_restore.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
It's only as an example.
Current version of (2.02.98-5) lvm2 allows you to i.e. 'export/swap' metadata from thin pool to a normal LV - and you could use then thin_check on a normal LV - fix metadata/ backup&restore them to bigger partition.
So I think the man page could be left as it is.
User should never try to fix/run think_check on 'running' thin pool.
The only exception currently is, when thin pool fails to start -
we leave the thin metadata device active - but it's rather 'temporary'
solution - and at this moment we should probably already disable this use,
since there is a way, to fix thin pool without this.
Obviously we are missing proper documentation how to fix thin pool metadata
with current lvm2 utilities, which needs to be written.
BZ#1022850 is a separate BZ for documenting fixing thin pool metadata with current lvm2 utilities.
There is now an lvmthin(7) man page which may address this issue.