HDDS-6228. Update config keys for open key cleanup service

[kaijchen]

What changes were proposed in this pull request?

Move config keys for open key cleanup service from OzoneConfigKeys to OMConfigKeys.
Change key names and default values as the design doc.

This change was part of #1511. The differences are:

1. org.apache.ratis.util.TimeDuration was replaced by java.time.Duration.
2. Instant.now() is only called once per query in getExpiredOpenKeys().

Since the service needs refinement such as Quota and FSO support,
let's update the config keys first.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-6228

How was this patch tested?

Integration tests.

[errose28]

Thanks for picking up this feature @kaijchen, and sorry I am so late on the review here. One minor inline comment but overall this looks good.

It does not look like the tests that were setting open key cleanup actually need the config, so maybe we can remove those settings in this PR too.

[errose28]

I think the Ratis TimeDuration was used since existing time configs in this class and the other config key classes use it. I am not sure if it provides any benefits over the Java Duration class, but I think we should keep these as TimeDurations for consistency.

[kaijchen]

Every other Ratis TimeDuration in this class is related to Ratis.
Either it starts with ozone.om.ratis.server or ozone.om.snapshot.
I changed this to java Duration because I think this config is not related to Ratis.
Do you think we still need to keep it as TimeDuration?

[errose28]

That's a good point. Actually it looks like most existing intervals are either configured as Strings with suffixes or as longs. If we want to be matching I think Strings would be the best bet, but Duration is fine too.

[kaijchen]

Yes, String is indeed better. And it's also the same representation as in ozone-default.xml.

[kaijchen]

Rebased to master to solve merge conflicts.

[captainzmc]

+1 Thanks @kaijchen for the patch. And thanks @errose28 for the review.

[kaijchen]

Thanks @errose28 and @captainzmc for the review.

It does not look like the tests that were setting open key cleanup actually need the config, so maybe we can remove those settings in this PR too.

Sorry I forgot to address this, I will fix it when implementing HDDS-4123.