You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When using lvconvert with --cachedevice with a large cache drive (~2TB), lvconvert will throw an error saying there are too many chunks and recommends using --chunksize to reduce the number.
% sudo lvconvert --type cache --cachedevice /dev/sdb vg/lv
Use all 1.63 TiB from /dev/sdb for cache? [y/n]: y
Creating cachevol LV lv with size 1.63 TiB.
Logical volume "lv_cache" created.
Cache data blocks 3506397184 and chunk size 128 exceed max chunks 1000000.
Use smaller cache, larger --chunksize or increase max chunks setting.
However, if you try to pass --chunksize to lvconvert it refuses to use it, making it impossible to proceed without increasing max chunks (which isn't desirable in this case).
% sudo lvconvert --type cache --chunksize 2M --cachedevice /dev/sdb vg/lv
Command does not accept option: --chunksize 2M.
Ideally, lvconvert should support --chunksize when combined with --cachedevice, or support automatic sizing like when using --cachepool
Unfortunately using --cachepool doesn't produce the same results as --cachepool expects free extents, which don't exist, as all of them have been allocated to either the origin LV or the cache.
% sudo lvconvert --type cache --cachepool lv_cache vg/lv
Using 2.00 MiB chunk size instead of default 64.00 KiB, so cache pool has less than 1000000 chunks.
WARNING: Converting vg/lv_cache to cache pool's data volume with metadata wiping. THIS WILL DESTROY CONTENT OF LOGICAL VOLUME (filesystem etc.)Do you really want to convert vg/lv_cache? [y/n]: y Volume group "vg" has insufficient free space (0 extents): 10 required. LV vg/lv_cache could not be converted to a cache pool.
The text was updated successfully, but these errors were encountered:
Background
Running on Debian Bullseye.
Problem
When using
lvconvert
with--cachedevice
with a large cache drive (~2TB),lvconvert
will throw an error saying there are too many chunks and recommends using--chunksize
to reduce the number.However, if you try to pass
--chunksize
tolvconvert
it refuses to use it, making it impossible to proceed without increasing max chunks (which isn't desirable in this case).Ideally,
lvconvert
should support--chunksize
when combined with--cachedevice
, or support automatic sizing like when using--cachepool
Unfortunately using
--cachepool
doesn't produce the same results as--cachepool
expects free extents, which don't exist, as all of them have been allocated to either the origin LV or the cache.The text was updated successfully, but these errors were encountered: