Open dhruvkb opened 3 years ago
Just an update, everything else seems to work normally if I comment out the service restart
operation in two places.
@dhruvkb which places? It's probable that the cavalcade implementation hasn't been updated in a very long time and that VVV has since switched to Ubuntu 18 and then 20 which use slightly different mechanism for services. It only works for you in the sense that cavalcade isn't running
You're right about Cavalcade not running. I just meant that commenting out Cavalcade allows the remaining provisioning process to complete, for anyone who was looking for a workaround.
It does seem like an Ubuntu version issue because now there's a different folder /etc/init.d
which I assume is also the reason cp
was failing.
Not sure if it's related but provisioning is now consistently failing with the message fatal: cannot change to '/etc/cavalcade': No such file or directory
.
Going by the logs the last thing to happen before that is the updating of WP-CLI plugins here: https://github.com/WordPress/meta-environment/blob/master/wordpressorg.test/provision/vvv-init.sh#L48
Given the final output is from WP CLI I expect that the cavalcade plugin is already installed and that, or something else, is attempting to trigger the Cavalcade runner which at that point isn't installed yet.
This PR might help https://github.com/WordPress/meta-environment/pull/158, part of the problem is the cavalcade service, and the installation of cavalcade to /etc/cavalcade
, I've made a minor refactor to try and handle the process better, but it requires testing. I'd be really appreciative if you could try it out and tell me if it helps and what the results are.
For reference, if the service fails to restart, it should now report that in the log rather than exiting
Initial testing run fails with the message fatal: destination path '/srv/www/wordpress-meta-environment/wordpressorg.test/public_html/wp-content/mu-plugins/cavalcade' already exists and is not an empty directory.
.
Figured that may be due to previous failed runs so ran a vagrant destroy && vagrant up
which unfortunately results in the same error message.
Then fully removed the www/wordpress-meta-environment
directory and provisioned again to ensure a truly clean environment and that does complete provisioning without failing.
It does however still show the message: Failed to restart cavalcade-wordpressorg.service: Unit cavalcade-wordpressorg.service not found.
Provisioning is no longer blocked, though the cavalcade service still needs fixing
The step involving copy of Cavalcade config
cavalcade-wordpressorg.conf
i.e.fails due to the directory structure not being present. The error message says "cp: cannot create regular file: No such file or directory". This can be fixed by a
mkdir
operation before thecp
operation (#154).With that change, the build goes a long way ahead but fails on the line
with the error message: "Failed to restart cavalcade-wordpressorg.service: Unit cavalcade-wordpressorg.service not found"