I was recycling every hour and/or 10000 requests I think. It did help extend the time in between crashes, but it didn't prevent them.
The only thing that seems definite is almost everyone was stable on 4.4.1. Somewhere in between there and 4.4.4, something got screwy.
The only thing that seems definite is almost everyone was stable on 4.4.1. Somewhere in between there and 4.4.4, something got screwy.
Comment