Drupal 10, the latest version of the open-source digital experience platform with even more features, is here.While running tests/functional_tests/aegir2-dev/run-tests.sh the provitioning fails to install hostmaster.
we use drush_set_default(), to set the document root for hostmaster... but this does not work when the process received a '--root=' (no value).
tests/functional_tests/aegir2-dev/modules/aegir/manifests/dev.pp
$platform_path = false,
...
if $platform_path { $o = " --root=${platform_path}" }I don't understand why the --root option gets added. In the meantime I've added a quickfix patch for provision to handle this.
| Comment | File | Size | Author |
|---|---|---|---|
| puppet-aegir-provision-quickfix.patch | 785 bytes | helmo |











Comments
Comment #1
helmo CreditAttribution: helmo commentedHmm, same error when the whole
is removed...
err: /Stage[main]/Aegir::Dev/Drush::Run[hostmaster-install]/Exec[drush-run:hostmaster-install]/returns: change from notrun to 0 failed: drush @none --yes --debug --working-copy --strict=0 --no-gitinfofile --aegir_version=6.x-2.x --script_user=aegir --aegir_root=/var/aegir --web_group=www-data --aegir_db_port=3306 --http_service_type=apache --aegir_host=aegir2-dev.test --http_port=80 hostmaster-install aegir2-dev.test >> /var/aegir/install.log 2>&1 returned 1 instead of one of [0] at /tmp/vagrant-puppet/modules-0/drush/manifests/run.pp:31Is there some magic to the 'root' word as option?
Comment #2
ergonlogicEvidently, --root shouldn't be passed by default. What does /var/aegir/install.log say is the problem?
Comment #3
helmo CreditAttribution: helmo commentedIt happily starts a drush make with an empty 'Aegir platform path' ... en eventually fails to copy the platform tree inplace. "Destination directory . is not writable. [44.12 sec, 6.14 MB]"
Comment #4
ergonlogicThe aegir2-dev test in puppet-aegir isn't passing a '--root' parameter. It's possible that #1635628: do not talk about a makefile if installing over an existing platform may have broken this. But then it's a core issue...
Comment #5
helmo CreditAttribution: helmo commentedI saw that commit while digging into this issue... but don'[t see how it could affect this.
Comment #6
ergonlogicIs this still an issue with the latest head?
Comment #7
helmo CreditAttribution: helmo commentedCan't reproduce it at the moment ... closing for now.