Correctly calculate directory sizes when using an object store primary storage#1630
Conversation
CarlSchwan
force-pushed
the
work/object-storage-compatibility
branch
from
ca51597 to
6a93479
Compare
CarlSchwan
force-pushed
the
work/object-storage-compatibility
branch
from
6a93479 to
c9e8ba0
Compare
PVince81
left a comment
PVince81
left a comment
There was a problem hiding this comment.
Alright, now I finally understand it. Makes absolute sense 👍
|
@CarlSchwan can you rebase ? it seems this repo has stricter rules to have up to date branches before merging. |
PVince81
added
4. to release
…y storage A object store storage should use an NoopScanner since it is not possible to scan the directory structure of an object store. This makes sure that the group folder mount point doesn't override the NoopScanner that the object store storage use. Signed-off-by: Carl Schwan <carl@carlschwan.eu>
CarlSchwan
force-pushed
the
work/object-storage-compatibility
branch
from
c9e8ba0 to
a9cf6da
Compare
|
/backport stable 22 |
|
/backport stable22 |
|
/backport stable22 |
|
/backport to stable22 |
|
👉 #1631 |
|
@Pytal I already created manual backport mr. Seems like I messed up the syntax for the backport bot |
|
@CarlSchwan would it be possible to backport this PR to v21 ? |
|
/backport to stable21 |
|
/backport to stable20 |
|
The backport to stable21 failed. Please do this backport manually. |
|
The backport to stable20 failed. Please do this backport manually. |
|
manual backport still required ? |
|
Hmm, tried to do it manually but it seems that it has already been done:
So, no ! :) |
6 participants
An object store storage should use a NoopScanner since it is not
possible to scan the directory structure of an object-store. This makes
sure that the group folder mount point doesn't override the NoopScanner
that the object store storage use.