In thWe have a need to have production cluster if I have a bug that is only reproducible in production,work machines (tin, terbium) to have HHVM (and thus Trusty). i can do something like:Use case is maintenance scripts that depend on HHVM.
sudo -u www-data hhvm -m debug /var/www/w/MWScript.php extensions/Flow/maintenance/foo.php --wiki=somewiki
and then set breakpoints, run the script, debug,etc.
NOTE: This also needs to be mirrored into Beta (somewhere, etc.
This is not currently possible onwe use deployment-bastionn like tin, as it does not have the hhvm executable. This is the same as tin in production. These are also not runnable from the deployment-mediawiki* servers as they do not have mwscript(likely same as mw* prod servers). In prod I would do this from terbium, so after talking to greg it seems like the logical choice is to spin up a new VM for beta that is the equivalent to terbiummaybe making a terbium-like in Beta?).