[DO NOT MERGE] Backport ETag improvements and bug fixes from master to ozone-1.4

[ivandika3]

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

[ivandika3]

@vtutrinov Could you help to check whether the ETag patches are correctly cherry-picked (especially HDDS-9680)?

[ivandika3]

cc: @xichen01

[ivandika3]

@adoroszlai I was hoping to add the ETag fix in #6519, but that might mean to exclude the s3a-test.sh during backport since it does not exist in ozone-1.4 branch yet. Do you think we should include this in this backport or a separate S3A compatibility backport?

[adoroszlai]

Thanks @ivandika3 for working on this.

Do you think we should include this in this backport or a separate S3A compatibility backport?

It can be included in a later PR. BTW, I just noticed a minor bug (HDDS-10719) introduced by HDDS-10615.

[vtutrinov]

@vtutrinov Could you help to check whether the ETag patches are correctly cherry-picked (especially HDDS-9680)?

@ivandika3 thanks for the backport, LGTM, +1

[ivandika3]

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