Add initial bcUpgradeTask

[mosche]

This can be run as follows until the incompatibility check is removed from TransportService

gradle -Dbwc.checkout.align=true -Dorg.elasticsearch.build.cache.push=true -Dignore.tests.seed -Dscan.capture-file-fingerprints -Dtests.bwc.main.version=9.1.0 -Dtests.bwc.refspec.main=e5cdc581cf7c03903a99eb2adec567294be9ad43 bcUpgradeTest -Dtests.jvm.argline="-Des.serverless_transport=true"

However, this sometimes failed unexpectedly on the upgrade due to an ML issue. When later trying to reproduce to see the actual exception, it didn't reproduce anymore :(

Failure running machine learning native code. This could be due to running on an unsupported OS or distribution, missing OS libraries, or a problem with the temp directory. To bypass this problem by running Elasticsearch without machine learning functionality set [xpack.ml.enabled: false].

Relates to ES-11870

[ldematte]

Failure running machine learning native code. This could be due to running on an unsupported OS or distribution, missing OS libraries, or a problem with the temp directory. To bypass this problem by running Elasticsearch without machine learning functionality set [xpack.ml.enabled: false].

I have seen this happening when the native controller does not start. This usually happens to me locally if there is already another controller running, e.g. because a node crashed (or I terminated/broken it in the debugger).
Could it be that we are not cleaning up things? Or just that you ctrl+c the job and left with a controller process hanging there?

[mosche]

I have seen this happening when the native controller does not start. This usually happens to me locally if there is already another controller running, e.g. because a node crashed (or I terminated/broken it in the debugger).
Could it be that we are not cleaning up things? Or just that you ctrl+c the job and left with a controller process hanging there?

You're probably right @ldematte , at some point I did clean up local process manually... not sure what caused them to fail in the first place, but afterwards this worked ok

[ldematte]

Looks good; I think this could be merged as-is and then the TODO will be addressed in a follow-up?

[mosche]

I think this could be merged as-is and then the TODO will be addressed in a follow-up?

Yes, that's my intention. Let's keep the initial test set as small as possible to not wait forever while working on this.

[elasticsearchmachine]

Pinging @elastic/es-core-infra (Team:Core/Infra)

[elasticsearchmachine]

💚 Backport successful

Status	Branch	Result
✅	9.0