Support OSD file backing #261
Conversation
Add support for auto-creating loopback backing for OSDs Signed-off-by: Peter Sabaini <[email protected]>
Signed-off-by: Peter Sabaini <[email protected]>
sabaini
force-pushed
the
feature/loopdev
branch
from
3373ec6 to
df2e26c
Compare
Thanks! I've now added longer help text for clarity |
| } | ||
|
|
||
| // Convert the size to MB. | ||
| switch strings.ToUpper(unit) { |
There was a problem hiding this comment.
Shouldn't this catch invalid units as well?
There was a problem hiding this comment.
The regexp should take care of that -- it matches only M, G or T. Anything else won't match and cause an "illegal spec" error
| // AddOSD adds an OSD to the cluster, given the data, WAL and DB devices and their respective | ||
| // flags for wiping and encrypting. | ||
| func AddOSD(s *state.State, data types.DiskParameter, wal *types.DiskParameter, db *types.DiskParameter) error { | ||
| logger.Debugf("Adding OSD %s", data.Path) | ||
|
|
||
| var err error | ||
|
|
||
| // sanity: loopback file and WAL/DB are mutually exclusive |
There was a problem hiding this comment.
is this a Ceph limitation? or something we don't support due to complexity in implementing it?
There was a problem hiding this comment.
My reasoning was that having loop files and using separate WAL/DB devices are orthogonal use cases, and we'd complicate implementation but also user interface surface without much benefit
Add support for auto-creating loopback backing for OSDs