[DO NOT MERGE] Backport ETag improvements and bug fixes from master to ozone-1.4
#6557
Conversation
…pache#5668) (cherry picked from commit 7370676)
(cherry picked from commit 5f6306d)
… key does not contain ETag field (apache#6377) (cherry picked from commit 3d193fc)
…tion (apache#6435) (cherry picked from commit c6c611f)
|
@vtutrinov Could you help to check whether the ETag patches are correctly cherry-picked (especially HDDS-9680)? |
|
cc: @xichen01 |
|
@adoroszlai I was hoping to add the ETag fix in #6519, but that might mean to exclude the |
|
Thanks @ivandika3 for working on this.
It can be included in a later PR. BTW, I just noticed a minor bug (HDDS-10719) introduced by HDDS-10615. |
@ivandika3 thanks for the backport, LGTM, +1 |
|
Thank you for the reviews @vtutrinov and @adoroszlai. Merged using: git checkout ozone-1.4
git pull
git merge --ff-only ozone-1.4-etag-backport
git push |
What changes were proposed in this pull request?
Backport the following ETag patches to ozone.1.4 for 1.4.1 release:
HDDS-9680. Use md5 hash of multipart object part's content as ETag (#5668)
HDDS-10395. Fix eTag compatibility issues during MPU (#6235)
HDDS-10403. CopyObject should set ETag based on the key content (#6251)
HDDS-10521. ETag field should not be returned during GetObject if the key does not contain ETag field (#6377)
HDDS-10587. Reset the thread-local MessageDigest instance during exception (#6435)
How was this patch tested?
https://github.com/ivandika3/ozone/actions/runs/8739462351