Ok. got the first breakthrough after all. I migrated SubQueries test.
Alan Woodward, you might be interested in noob's feedback.

Copying config folder under the given name to Zookeeper, is not obvious at all:

    String configName = "solrCloudCollectionConfig";
    int nodeCount = 5;
    configureCluster(nodeCount)
       .addConfig(configName, configDir)
       .configure();

if it's possible, I ask to simplify it if possible, perhaps more descriptive names or perhaps implicit behavior, or so.
Also, I experienced a typical leg-shooting: the test had solrconfig-basic.xml w/o update log. Somewhere deep in log it was reported somehow like ERROR ... RecoveryStrategy No UpdateLog found cannot recover. It's not easier to figure out the reason looking at log tail. So, ideally there should be a circuit-breaker preventing cloud launch without updatelogs.
These two thoughts is just FYI, not really need to act.

Copying config folder

Agreed, this is a bit of a pain at the moment. You need to do something like {{{TEST_PATH().resolve("configsets").resolve(name).resolve("conf") - I'll add a method to SolrCloudTestCase to abstract this away.

So, ideally there should be a circuit-breaker preventing cloud launch without updatelogs

This would be a nice to have, yes! There are a few things that are required, both in solrconfig and schema, for cloud to work, and it would be good to have some kind of checking at core-load time - do you want to open a separate JIRA for that?

spawned SOLR-9130

Alan Woodward, is it worth to migrate non-Zk test {{BlockJoinFacetDistribTest extends BaseDistributedSearchTestCase }} ?

The main reason I have for migrating the tests is to get off the legacy cloud formats, which don't apply to anything built on BaseDistributedSearchTestCase, so it's up to you really. You may find the test runs faster, as MiniSolrCloudCluster starts up its cores in parallel.

moved DistribJoinFromCollectionTest.
One detail:

• collection "to" is 2 shards x 2 replicas
• collection "from" is 1 shard x 4 replicas, not fewer.
  Timothy Potter, do you have any concerns with this amendments for SOLR-4905?

migrated the 3rd test.
One question: is it really necessary to mention tests only commit at CHANGES.txt ?

One question: is it really necessary to mention tests only commit at CHANGES.txt ?

It isn't necessary, but it can be quite helpful, since CHANGES.txt is often used as a way of understanding the project. For example, for the 6.0.1 release, I'm using CHANGES.txt to decide which issues to backport.

If we're more careful in the future with JIRA fix version setting, it's possible that CHANGES.txt could be auto-generated from JIRA - I know some projects do that, and it's been discussed in the past for Lucene/Solr.

it seems ready. But yesterday I've got non-reproducible failure on BlockJoinFacet test. Will check some more time.

fixed failures by clearing indexes. I'm going to commit it quite soon.

Commit ffa2d65efe5ed2eea04cfe60a6d97828c98da7b0 in lucene-solr's branch refs/heads/master from Mikhail Khludnev
[ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=ffa2d65 ]

SOLR-9110: moving JoinFromCollection- SubQueryTransformer- BlockJoinFacet- Distrib Tests to SolrCloudTestCase

Commit aec3654fb8e6530494d909d57dada19642db1128 in lucene-solr's branch refs/heads/branch_6x from Mikhail Khludnev
[ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=aec3654 ]

SOLR-9110: moving JoinFromCollection- SubQueryTransformer- BlockJoinFacet- Distrib Tests to SolrCloudTestCase

attaching test fix: removing static refs.

Commit 2391b5b81dfade413d70d26da63e85cbd7245bf2 in lucene-solr's branch refs/heads/master from Mikhail Khludnev
[ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=2391b5b ]

SOLR-9110: removing static members from tests.

Commit 9863eea256c649dd035103128f95410bd85b8771 in lucene-solr's branch refs/heads/branch_6x from Mikhail Khludnev
[ https://git-wip-us.apache.org/repos/asf?p=lucene-solr.git;h=9863eea ]

SOLR-9110: removing static members from tests.