問62
頻出度★★☆
複数のハードディスクを論理的に一つのものとして取り扱うための方式①～③のうち, 構成するハードディスクが 1 台故障してもデータ復旧が可能なものだけを全て挙げたものはどれか。 ① RAID5 ② ストライピング ③ ミラーリング
平成31年度 春期 ITパスポート試験 問62
テクノロジ系システム構成要素 › システムの構成
This question's explanation and your answers so far.
解説
問 62
Incorrectア
“(1) and (2)”. Striping (RAID0) in (2) merely splits the data and writes it in parallel across multiple disks; it has no redundancy, so a single disk failure loses all data and cannot be recovered. Including (2) already makes this wrong.
Incorrectイ
“(1), (2), and (3)”. (1) and (3) are correct, but this also includes striping (2), which cannot be recovered. It violates the condition of listing only the recoverable methods, so it is wrong.
- ウ
正解 Only a method that has redundancy (duplicated data or error-correcting information) can recover data when one disk fails. RAID5 (1) records the data together with parity (error-correcting information) spread across multiple disks, so even if one disk fails, the original data can be recomputed and restored from the remaining disks and the parity. Mirroring (3, RAID1) writes the same data to two disks at once, so if one breaks, the data can still be read from the other. Therefore only (1) and (3) can recover, and option ウ, which lists exactly these two, is correct.
Incorrectエ
“(2) and (3)”. Mirroring (3) is correct, but striping (2) has no redundancy and cannot be recovered. It also drops RAID5 (1), which is recoverable, so it fails the condition on two counts.
ここがポイント
- RAID fault tolerance: RAID5 spreads parity (error-correcting information) across disks and mirroring (RAID1) duplicates the data, so both can recover data after one disk fails.
- Striping (RAID0) aims for speed through split parallel writes but has no redundancy; losing all data when a single disk fails is the decisive difference from the other methods.
用語