Clean already exist tmp directory when use s3 rest api when initiate multipart upload use s3 rest api.

According to the description in the aws s3 document, the s3 service will only store the latest data (that is, directly overwrite the existing data). The work of checking whether the data already exists should be done by the client.

Amazon S3 is a distributed system. If it receives multiple write requests for the same object simultaneously, it overwrites all but the last object written. Amazon S3 does not provide object locking; if you need this, make sure to build it into your application layer or use versioning instead. PutObject

Therefore, many third-party clients do not support the process of first checking and then overwriting (like spark and mingo).

#14203 In this pr, @ZacBlanco already implement overwrite in createObjectOrUploadPart func.

So, do we also need to do some cleanup operations when initializing multipart upload the request? For example, delete the tmp directory and the multipart upload file.

[jffree]

@ZhuTopher #14487 This pr (not merged) is used to clean up the tmp directory when user fails to upload the file and then retry the upload. #14328 This pr (merged) is used to clean up the tmp directory when user fails to upload the file and does not retry the operation.

[ZhuTopher]

@jffree I recall you had a PR with these changes already, or did I misremember that? Either way I’m in agreement with:

There are two scenarios here: one is that the user fails to upload the file and does not retry the operation. At this time, the tmp directory should be cleaned by lazy scanning (#14328 ). The other is that the user fails to upload the file and then retry the upload. At this time, the tmp directory should be cleaned up immediately, without affecting the retry operation.