HDDS-2426. Support recover-trash to an existing bucket.

[cxorm]

What changes were proposed in this pull request?

This PR composed of changes including:

• provide trash-enabled config when creating buckets.
• provide recover-window config when creating buckets.
• global config of above settings.
• core logic of recovering trash to existing buckets.

Proposed fix.

• In OmMetadataManagerImpl#getPendingDeletionKeys we only check last item in repeatedOmKeyInfo.
  Cause once the last item is trash-enabled, we keep this key from purging in OM DB.

  If the life time of key in deletedTable is less than recoverWindow, we didn't purge it in this time.
  So we use the difference between currentTime and modificationTime of key to check the situation.

• Using trashTable to track trash.

  In part of deleting key,
  add cache-update of trashTable in OMKeyDeleteRequest and DB-operation of trashTable in OMKeyDeleteResponse

  In part of recovering trash,

  for OMTrashRecoverRequest
  remove OMKeyInfo in RepeatedOmKeyInfo from cache of trashTable and deletedTable.
  (not remove all RepeatedOmKeyInfo.)
  add OMKeyInfo to cache of keyTable.

  for OMTrashRecoverResponse
  remove OMKeyInfo in RepeatedOmKeyInfo of trashTable and deletedTable in OM DB.
  add OMKeyInfo to KeyTable in OM DB.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-2426

How was this patch tested?

Add UTs to track the OMTrashRecoverRequest and OMTrashRecoverResponse.

[cxorm]

Rebased latest master-branch (#954 ) and trigger CI

[cxorm]

I'm going to fixing violation.

[cxorm]

This violation is not related to the change.
If anyone think it is improper, we could trigger actions again.
(I have test it with the same change here, and it passed all checks)

[bshashikant]

Thanks @cxorm for working on this. I have started the review and will have a deeper look into it.
@bharatviswa504 , can you plz review this as well?

[bshashikant]

its better to name it as isTrashEnabled()?

[bshashikant]

rename as isTrashEnabled()?

[bshashikant]

can we rename the function shouldDelete() ? the name looks confusing

[bshashikant]

why do we need to update the cache for both trash and deleted table here?

[bshashikant]

Ideally, if a key gets deleted from a bucket which is trash enabled, i think we should just add it to the trash table, not the deleted table, otherwise it will lead to having duplicate entries in both trash and delete table. Once the trash interval expires, keys can be moved from trash table to deleted table using a background task which will then be cleaned by the KeyDeleting Service. What do you think?

[cxorm]

Thank you @bshashikant for the review.
I'm working on this these days, and I would draw a picture for this question.

[cxorm]

Ideally, if a key gets deleted from a bucket which is trash enabled, i think we should just add it to the trash table, not the deleted table, otherwise it will lead to having duplicate entries in both trash and delete table. Once the trash interval expires, keys can be moved from trash table to deleted table using a background task which will then be cleaned by the KeyDeleting Service. What do you think?

Yeah, I agree that we shouldn't have duplicated entries in both and delete table.
This part would be updated .

[bshashikant]

Let's hold off for now on this. There are some ideas being discussed around this.

[cxorm]

Yeah, thanks @bshashikant for the remind.

[cxorm]

I am going to take it this week.

[cxorm]

Rebased latest master-branch (#1109 ) and fixed conflict.

[bshashikant]

/pending

[github-actions]

Marking this issue as un-mergeable as requested.

Please use /ready comment when it's resolved.

/pending

[github-actions]

Thank you very much for the patch. I am closing this PR temporarily as there was no activity recently and it is waiting for response from its author.

It doesn't mean that this PR is not important or ignored: feel free to reopen the PR at any time.

It only means that attention of committers is not required. We prefer to keep the review queue clean. This ensures PRs in need of review are more visible, which results in faster feedback for all PRs.

If you need ANY help to finish this PR, please contact the community on the mailing list or the slack channel."