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