Stewart X Addison

Ref https://github.com/adoptium/temurin-build/issues/2865

@gdams As per the [slack thread](https://adoptopenjdk.slack.com/archives/C53GHCXL4/p1596121051329700), can you look at this please? Not sure where the logic is that determines how the mapping is defined for the `dockerBulid` host situation

All of the docker build images should set them directly e.g. https://github.com/adoptium/infrastructure/blob/69b696162f37caf612ae7a035c4a8659a64f8f3a/ansible/docker/Dockerfile.Alpine3#L16 so I can't see why it would make more sense to pull the variable from the host in...

@smlambert Of course the obvious concern with this is that with [HotSpot debug images being added soon](https://github.com/AdoptOpenJDK/openjdk-build/pull/2443) it will be even more of an increase to our space requirements :-)

Running https://ci.adoptopenjdk.net/job/build-scripts/job/utils/job/pipeline_jobs_generator_jdk11u/89/console has generated the job although it [failed](https://ci.adoptopenjdk.net/job/build-scripts/job/jobs/job/jdk11u/job/jdk11u-linux-x64-fast_startup/1/console) due to https://github.com/adoptium/infrastructure/issues/2491

Leaving this open as we should determine why the log is claiming to be generating this jobs, but is not doing so.

New related issue: systemtests on OpenJ9 have started running slowly over the weekend - raised at https://github.com/eclipse/openj9/issues/11080

Since we seem to be sprawling issues here: - Determing test schedule: https://github.com/AdoptOpenJDK/openjdk-tests/issues/2037 - JDK8: https://github.com/AdoptOpenJDK/openjdk-tests/issues/2050 - JDK11: https://github.com/AdoptOpenJDK/openjdk-tests/issues/2051

@johnoliver are you able to look into this? Is it just because of all the locked GA builds being retained?

Current numbers: ``` ==== openjdk10-pipeline 4 4 120 ==== openjdk11-pipeline 47 47 1410 ==== openjdk12-pipeline 33 33 990 ==== openjdk13-pipeline 24 24 715 ==== openjdk8-pipeline 47 47 1363 ==== openjdk9-pipeline...