HDDS-9051. Change the log level when no nodes are available in NetworkTopologyImpl

[Tejaskriya]

What changes were proposed in this pull request?

We currently have a log message in NetworkTopologyImpl set at the WARN level for when no datanodes are available. It’ll usually show up with a bunch of other related, repeating logs. It is better to set it to DEBUG level, and relying on the INFO message at the end as a summary of what happened.

What is the link to the Apache JIRA

https://issues.apache.org/jira/browse/HDDS-9051

How was this patch tested?

Only log level change. No testing required.

[tanvipenumudy]

Thank you @Tejaskriya for the patch.

It is better to set it to DEBUG level, and relying on the INFO message at the end as a summary of what happened

Can you please specify the INFO message which should suffice instead of the WARN messages?

[Tejaskriya]

Can you please specify the INFO message which should suffice instead of the WARN messages?

The method chooseNodeInternal() in which this log is present is finally being called by SCMContainerPlacementRackAware#chooseNode() and SCMContainerPlacementRackScatter#chooseNode() . Both these methods have Log.info() messages to log "No satisfied datanode" being found after maxTries are hit. These INFO logs will serve the purpose of showing us the summary

[siddhantsangwan]

@Tejaskriya The purpose here is to remove redundant information in the info logs. To do this, we need to ensure all classes using this method from NetworkTopology are providing a summary or similar information so that we can get rid of the repeated warnings from this class. SCMContainerPlacementRackScatter already does this. What about other classes?

[siddhantsangwan]

Thinking about it some more, it's actually hard to guarantee what a future user of the class will do, and the implementation details of this class cannot really rely on that. I think we're better off just bringing down the log level from WARN to INFO.

[Tejaskriya]

I think we're better off just bringing down the log level from WARN to INFO.

We want to bring it down from WARN to INFO and not DEBUG?

[siddhantsangwan]

We want to bring it down from WARN to INFO and not DEBUG?

Yes

[Tejaskriya]

@siddhantsangwan thank you for the review! I have changed the log level to INFO now. Could you please approve the workflows if the patch is good to go?

[siddhantsangwan]

Thanks for the update, I've started CI.

[adoroszlai]

Thanks @Tejaskriya for the patch, @siddhantsangwan, @tanvipenumudy for the review.