RAID controller 0 has failed in the SATABoy that is connected to FW1, as evidenced by the audible alarm and serial console log messages.  Looking at the uptime trend in DataViewer, it looks like FW1 crashed at ~02:23 UTC on 8 Sep and did not start (reliably) running again until 06:11 UTC the same day.  That range probably corresponds to the time when the controller failed, or started to fail.  Controller 1 appears to be active and working at present.  For now I've silenced the audible alarm until Dan can hopefully come up and have a look at it.