it's obvious in retrospect, but I was recently surprised to find how huge a difference it makes to use larger blocks with raid, especially raid5. (a large, aligned r5 write can avoid parity reads, and thus *flies*!). so how do you get the filesystem to produce larger IOs? "large" here is much larger than the typical FS blocksize. ideally, you want chunks of, say, 64K going to each disk, so if you have an 8+1 r5, you'd really like to see writes of at least 512K. I was wondering whether perhaps DM could "reblock" a filesystem. thanks, mark hahn. - 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
