HDDS-15481. Avoid re-encoding the URL safe token for each read chunk

[sodonnel]

What changes were proposed in this pull request?

This is a small performance improvement - the Token is re-encoded as a URL-safe base64 string on every call to readBlockImpl(), which will be called for each 1MB read.

We can cache the URL encoded value and on re-encode it if the token is refreshed.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-15481

How was this patch tested?

Depends on existing tests.

[szetszwo]

@sodonnel , this is a good catch! This problem was copied from BlockInputStream. How about we also fix BlockInputStream here? We should just change the token class from Token<OzoneBlockTokenIdentifier> to String in all the places.

[sodonnel]

We should just change the token class from Token to String in all the places.

How far do you think we should go? Token is used all over the place - intellij just gives up searching for usages.

The nicest fix would be to cache the encoded token inside the token object itself, but its a hadoop class, so we cannot easily change it just in Ozone.

[szetszwo]

Just BlockInputStream/StreamBlockInputStream. I tried to change the code and the change seems not big. The reason to involve BlockInputStream here is that they have some shared code.

[github-actions]

This PR has been marked as stale due to 21 days of inactivity. Please comment or remove the stale label to keep it open. Otherwise, it will be automatically closed in 7 days.

[github-actions]

Thank you for your contribution. This PR is being closed due to inactivity. Please contact a maintainer if you would like to reopen it.