We need to convert the jobrunner cluster to HHVM. This is the tracking ticket for this activity.
Description
Description
Event Timeline
Comment Actions
First, we're reimaging an appserver so that we don't lower the throughput of jobrunners if something goes horribly wrong
Comment Actions
The reimaged server, mw1152, is performing very well, it actually has more than doubled the throughput compared to "normal" jobrunners, and it looks stable.