LCG_94 Update, 1.0 branch (2018.09.18.)

This is the "1.0 version" of !338 (merged). I had to manually cherry-pick changes from that MR, as in 1.0 we only want to upgrade AthAnalysisExternals to LCG_94.

I also included updates to all of the packages used by AnalysisBaseExternals, so that they would build versions of everything that are compatible with LCG_94.

This MR may stay open for a bit, while I'm doing some tests in 21.0 and 21.2 with these updates... Also note that the updates to the standalone externals will have to be cherry-picked into master as well eventually.

AE Build FAILURE
Build logfiles are available at Jenkins [AE-MERGE-REQUEST #540]

AE Build FAILURE
Build logfiles are available at Jenkins [AE-MERGE-REQUEST #541]

The first error was very clearly a glitch in Jenkins. But the second one is just super weird. I opened this discussion about it:

https://root-forum.cern.ch/t/very-strange-build-error-in-root-6-14-04/

For now I'll just retry the build once more...

AE Build FAILURE
Build logfiles are available at Jenkins [AE-MERGE-REQUEST #542]

Okay, time to call in some help. @aundrus, do you know what's going on with Jenkins? 2 out of these 3 attempts died with Jenkins authentication errors like:

[EnvInject] - Unset unresolved 'BASH_FUNC_module()' variable.
 > /cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase/x86_64/git/2.11.1-x86_64-slc6/bin/git rev-parse --is-inside-work-tree # timeout=10
Fetching changes from 2 remote Git repositories
 > /cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase/x86_64/git/2.11.1-x86_64-slc6/bin/git config remote.atlasexternals.url ssh://git@gitlab.cern.ch:7999/akraszna/atlasexternals.git # timeout=10
Fetching upstream changes from ssh://git@gitlab.cern.ch:7999/akraszna/atlasexternals.git
 > /cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase/x86_64/git/2.11.1-x86_64-slc6/bin/git --version # timeout=10
using GIT_SSH to set credentials username with pkey from file full path
 > /cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase/x86_64/git/2.11.1-x86_64-slc6/bin/git fetch --tags --progress ssh://git@gitlab.cern.ch:7999/akraszna/atlasexternals.git +refs/heads/*:refs/remotes/atlasexternals/*
ERROR: Error fetching remote repo 'atlasexternals'
hudson.plugins.git.GitException: Failed to fetch from ssh://git@gitlab.cern.ch:7999/akraszna/atlasexternals.git
	at hudson.plugins.git.GitSCM.fetchFrom(GitSCM.java:803)
	at hudson.plugins.git.GitSCM.retrieveChanges(GitSCM.java:1063)
	at hudson.plugins.git.GitSCM.checkout(GitSCM.java:1094)
	at hudson.scm.SCM.checkout(SCM.java:495)
	at hudson.model.AbstractProject.checkout(AbstractProject.java:1276)
	at hudson.model.AbstractBuild$AbstractBuildExecution.defaultCheckout(AbstractBuild.java:560)
	at jenkins.scm.SCMCheckoutStrategy.checkout(SCMCheckoutStrategy.java:86)
	at hudson.model.AbstractBuild$AbstractBuildExecution.run(AbstractBuild.java:485)
	at com.tikal.jenkins.plugins.multijob.MultiJobBuild$MultiJobRunnerImpl.run(MultiJobBuild.java:136)
	at hudson.model.Run.execute(Run.java:1735)
	at com.tikal.jenkins.plugins.multijob.MultiJobBuild.run(MultiJobBuild.java:73)
	at hudson.model.ResourceController.execute(ResourceController.java:97)
	at hudson.model.Executor.run(Executor.java:405)
Caused by: hudson.plugins.git.GitException: Command "/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase/x86_64/git/2.11.1-x86_64-slc6/bin/git fetch --tags --progress ssh://git@gitlab.cern.ch:7999/akraszna/atlasexternals.git +refs/heads/*:refs/remotes/atlasexternals/*" returned status code 128:
stdout: 
stderr: Permission denied (publickey).
fatal: Could not read from remote repository.

And even on the actual build error I'm very suspicious that it was aibuild092 that messed something up. (@obreshko and @boeriu are of course very welcome to have a look as well if they have ideas what to look for.)

Just to confirm that we are aware of that issue and @aundrus is following up on it. There is also "CI failures after Reboot of build nodes." discussion at hn-atlas-SITinternal@cern.ch

AE Build FAILURE
Build logfiles are available at Jenkins [AE-MERGE-REQUEST #543]

added 1 commit

• acb49c5e - Forcing ROOT to build VecCore itself.

Compare with previous version

See

https://root-forum.cern.ch/t/very-strange-build-error-in-root-6-14-04/

for some further info on why I (now) think that the ROOT build kept failing in the CI. Hopefully the upcoming build will now finally succeed...

AE Build SUCCESS
Build logfiles are available at Jenkins [AE-MERGE-REQUEST #544]

Okay... Now the only worrisome thing is that there's a chance that when a CI machine switches back to ROOT 6.12/06 after building this update, that may also fail. Though maybe not...

At this point I'm running my final test with Athena (on top of LCG_88). If that succeeds as well, this can go in.

merged

mentioned in commit f2d53a6c

mentioned in merge request !343 (merged)