Just thought I would expand on the SOAP thing. SOAP is not the solution:
SOAP uses spectrography to detect microscopic metallic particles such as are produced by general wear, plain bearing degradation etc
In the REDL case, the problem seems to be that a chunk of roller bearing raceway surface broke away due to spalling. This left a pit which was the likely start of a fatigue crack. Neither the chunk, or the fatigue crack, released any microscopic wear particles, so there would be nothing for SOAP to detect.
Therefore SOAP or its ilk would not have shown any problem prior to the accident, and in this respect unfortunately the otherwise good AAIB report is misleading (aka incorrect).
Chip detection is the only current means to detect this type of failure. It did of course work correctly, but unfortunately the implications of the chip were not correctly grasped by the maintenance guys and the accident happened. Silver and cadmium are non-magnetic so it seems pretty unlikely to me that they would be stuck to magnetic chip detector!