Outage at the Old Reader. “When all your engines stop your flight is over”
Multiple SSD fails in new storage deployment. Are SSDs more vulnerable to fails? Is there a possible useful burn-in procedure? Given the bathtub curve, we expect a higher rate of fails for new devices - any good ideas to minimise risk from that? Arranging for storage devices to come from different vendors and different batches is good practice. Choosing a RAID architecture which will survive multiple fails and support rapid rebuilds will help.