Cherry-pick "Merge branch 'DoxygenRevival-master-20180917' into 'master'" into 1.0

Doxygen Target Revival, master branch (2018.09.19.)

See merge request !342 (merged)

(cherry picked from commit b9a699ca)

• ef254dee Revived the "doc" target for ATLAS projects.
• 82cd70c6 Protected the Doxygen code against directory names with spaces in them.

AE Build FAILURE
Build logfiles are available at Jenkins [AE-MERGE-REQUEST #549]

@aundrus, @obreshko, is there some new issue?

13:48:07 Using default cache: /tmp/krb5cc_97155_TAolsY
13:48:07 Using principal: atlasbot@CERN.CH
13:48:07 Using keytab: /var/lib/jenkins/atlasbot.keytab
13:48:07 Authenticated to Kerberos v5
13:48:07 klist: No credentials cache found (ticket cache FILE:/tmp/krb5cc_97155_TAolsY)
13:48:07 eosfusebind: could not find automatically a credential file
13:48:07 + kinit -R
13:48:07 kinit: Credentials cache file '/tmp/krb5cc_97155_TAolsY' not found while renewing credentials
13:48:07 [EnvInject] - Script executed. The exit code is 1.
13:48:07 [EnvInject] - Fail the build.
13:48:09 Warning: you have no plugins providing access control for builds, so falling back to legacy behavior of permitting any downstream builds to be triggered
13:48:09 Returning node parameter for aibuild092
13:48:09 Triggering a new build of AE-post-results
13:48:09 Finished: FAILURE

I thought the authentication issues should be fixed for the moment...

AE Build SUCCESS
Build logfiles are available at Jenkins [AE-MERGE-REQUEST #550]

Well, the retry worked. So it was probably just a transient glitch...

merged

mentioned in commit 730fd771

@akraszna : this looks like a file system glitch to me . . . the kerberos cache file was not found. The subsequent job found it successfully:

14:10:52 Using default cache: /tmp/krb5cc_97155_TAolsY

... hopefully one of those spurious things that happens once a year or once an ATLAS lifetime.

AE job configuration include post-processing step "AE-post-results". The post processing step from the previous job indicates simultaneous running kinit command: http://aibuild080:8080/job/AE-post-results/62/console:

13:48:07 + kinit atlasbot@CERN.CH -5 -V -k -t /var/lib/jenkins/atlasbot.keytab 13:48:07 Using default cache: /tmp/krb5cc_97155_TAolsY 13:48:07 Using principal: atlasbot@CERN.CH 13:48:07 Using keytab: /var/lib/jenkins/atlasbot.keytab 13:48:07 Authenticated to Kerberos v5

Two simultaneous kinit commands (re)creating the same cache file is not a good thing. Perhaps I need to review AE configurations (still practically unchanged since Christian designed them) for removing the possibility of this.

mentioned in commit fwinkl/athena@99658133

mentioned in merge request athena!20441 (merged)

mentioned in commit fwinkl/athena@1b6e5461