On 12/17/2013 05:40 AM, Nikolaus Jeremic wrote: > Hi, > > I've did some Linux MD RAID 5 and 6 random write performance tests with > fio 2.1.2 (Flexible I/O tester) under Linux 3.12.4. However, the results > for RAID 6 show that writes to a single chunk in a stripe (chunk size is > 64 KB) result in more than 3 reads in case of more than 6 drives (tested > with 7, 8, and 9 drives) in the array (see fio statistics below). It > seems like that in the event of updating one data chunk in a stripe, all > of the remaining data chunks are read. > > By the way, in case of RAID 5 and 5 or more drives, the remaining chunks > seem not to be read when updating a single chunk in a stripe. This is not a bug. When writing to a small part of a stripe, the parity must be recomputed for the whole stripe, causing MD to read the rest of the stripe. However, it is mathematically possible to compute the new parity given the new data, old data, and the old parity. This is a simple computation for raid5 and this shortcut has been implemented. The similar shortcut computation for raid6 has been discussed, but no-one has provided a patch. (It is not so simple.) I suspect a patch would be welcome. :-) HTH, Phil -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
