Comment on Race-condition-free deployment with the "symlink replacement" trickparentComments−dvirsky13yI've been using this pattern for almost 2 years, and indeed it presented a problem at first with bytecode cache. We've added a simple post deploy script that flushes APC and it fixed this.
Comments
I've been using this pattern for almost 2 years, and indeed it presented a problem at first with bytecode cache. We've added a simple post deploy script that flushes APC and it fixed this.