Board Thread:Questions and Answers/@comment-9941439-20131203191949/@comment-1757231-20131215040100

It seems like somebody used a deployer to use a REP. REP's are terrible, and they are the "best" way to crash a server. I also suggest not using 3.1.2, but 3.1.3.

In the end, fix the crash by setting the ID of REP (from WR-CBE) to 0/-1. Then, future problems could be avoided.