1==================
2Partial Parity Log
3==================
4
5Partial Parity Log (PPL) is a feature available for RAID5 arrays. The issue
6addressed by PPL is that after a dirty shutdown, parity of a particular stripe
7may become inconsistent with data on other member disks. If the array is also
8in degraded state, there is no way to recalculate parity, because one of the
9disks is missing. This can lead to silent data corruption when rebuilding the
10array or using it is as degraded - data calculated from parity for array blocks
11that have not been touched by a write request during the unclean shutdown can
12be incorrect. Such condition is known as the RAID5 Write Hole. Because of
13this, md by default does not allow starting a dirty degraded array.
14
15Partial parity for a write operation is the XOR of stripe data chunks not
16modified by this write. It is just enough data needed for recovering from the
17write hole. XORing partial parity with the modified chunks produces parity for
18the stripe, consistent with its state before the write operation, regardless of
19which chunk writes have completed. If one of the not modified data disks of
20this stripe is missing, this updated parity can be used to recover its
21contents. PPL recovery is also performed when starting an array after an
22unclean shutdown and all disks are available, eliminating the need to resync
23the array. Because of this, using write-intent bitmap and PPL together is not
24supported.
25
26When handling a write request PPL writes partial parity before new data and
27parity are dispatched to disks. PPL is a distributed log - it is stored on
28array member drives in the metadata area, on the parity drive of a particular
29stripe.  It does not require a dedicated journaling drive. Write performance is
30reduced by up to 30%-40% but it scales with the number of drives in the array
31and the journaling drive does not become a bottleneck or a single point of
32failure.
33
34Unlike raid5-cache, the other solution in md for closing the write hole, PPL is
35not a true journal. It does not protect from losing in-flight data, only from
36silent data corruption. If a dirty disk of a stripe is lost, no PPL recovery is
37performed for this stripe (parity is not updated). So it is possible to have
38arbitrary data in the written part of a stripe if that disk is lost. In such
39case the behavior is the same as in plain raid5.
40
41PPL is available for md version-1 metadata and external (specifically IMSM)
42metadata arrays. It can be enabled using mdadm option --consistency-policy=ppl.
43
44There is a limitation of maximum 64 disks in the array for PPL. It allows to
45keep data structures and implementation simple. RAID5 arrays with so many disks
46are not likely due to high risk of multiple disks failure. Such restriction
47should not be a real life limitation.
48