問題文
Two accelerators in the same server both report remapped rows. On the first, the reserved-row availability is still reported as high and no remapping has ever failed. On the second, availability is reported as none and a remapping failure has been recorded in the past. The team has one spare card. Which device should be replaced and why?
選択肢
- Neither, and remapping is a normal maintenance activity of the memory subsystem while a recorded failure simply means the operation was retried successfully on a different bank later.
- The second one, because it has exhausted the reserved rows it can use and has already failed a remap, so further defects cannot be absorbed.
- The first one, and any device that reports remapped rows at all has already started to degrade while the number of reserved rows still available does not change that conclusion in any way.
- Both, since the server should be replaced as a unit.