I am currently creating a raid 6 array, normally I would just let the defaults do the job and be done with it but I'm having to do a limited create and then add disks. >From what I can recall a raid six can have differing layouts, the "checksum" data can be interspersed across the disks or put on 2 "parity" disks and the data contained on the other set of drives. My first question is, does this still hold true and if so do I need to do anything to get the parity spread over the disks, is there a layout recommended or is the "if not specified it will do this" route possible. My second question is, due to my situation I am going to have to create a 4 drive raid 6 array with just 2 disks initially. The reason for this is a lack of sata ports/disks (I have the 4 drives, but... the end partition of 2 of the drives is needed to keep a raid 6 6 disk set up running, I will drop 2 drives/partitions from the 6, extend the lower portion of the 2 drives (new raid 6), then copy from the degraded 6 drive to the new degraded 4 drive, then delete partitions/disks from the 6 drive and extend the last 2 drives, of the 4 drives, and add them in.... Now my thinking is that all I need to do is, fail/remove "drive b/partition6 from the six drive, then drive c/partition6... I now have a degraded raid 6. Then create MD51 /dev/sdb5 /dev/sdc5 missing missing. Copy data to md51 (after file system creation) add MD51 /dev/sdd5 /dev/sde5 and jobs a good un. My issue, I may be over thinking this, is that the b5/c5/missing/missing might force the d5/e5 to be parity only, when I would like the parity to be across all drives... a nice clean stripe! Or perhaps it makes no difference to performance/disk life/summat else. Jon -- 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
