Example: Writing to a failed disk unit

A write operation from the iSeries™ server detects that the disk unit that is to contain the data has failed. The write operation is to disk unit 3, sector 1. The following actions occur:

  1. The original data is lost on disk unit 3, sector 1, because of the failure.
  2. The new parity data is calculated by reading disk unit 1, sector 1, and disk unit 2, sector 1.
  3. New parity information is calculated.
  4. New data cannot be written to sector 1 on disk unit 3 because of the failure.
  5. New parity information is written to parity sector 1 on disk unit 4.

    Write operations require multiple read operations (n-2 reads, where n is the number of disk units) and only one write operation for the new parity information. Data from disk unit 3 will be rebuilt during synchronization after disk unit 3 is replaced.