> Tom> is totally irrational for a general default. I mean, given that it > Tom> was 1024 (512k), try to double it? Fine. Try to quadruple it? > Tom> Alright. We'll need to deal with some alignment / boundary issue > Tom> (like the typical 65535 vs 65536 case)? Okay let's do it. But > Tom> what's the sense in picking a random RAID configuartion as the base > Tom> to decide the default? > > I agree that the new default is completely arbitrary. And I think there > was a bit of controversy at the time. However, the numbers were > compelling so the change went in. For older SCSI and especially ATA drives (and it wouldn't surprise me if it is true of modern ones) there are also huge latency tradeoffs. Before you jump up and down about numbers what are the latency numbers like on classic ATA drives with that sized block I/O. You could easily up your RAID numbers while wrecking realtime and desktop performance. Without NCQ you certainly don't want huge I/O's blocking up the controller, with NCQ it depends if the firmware these days has improved and understands any kind of fairness or like the early ones and older SCSI drives will happily starve requests. Conceptually it also seems wrong - if you want to avoid partial stripe writes on a consumer grade disk subsystem then use smaller stripes. At the ATA level we can detect both the presence of command queueing ability, and also whether the device is spinnning rust or not, so it may be smarter defaults could be done based upon whether the device is an SSD or not. Alan -- To unsubscribe from this list: send the line "unsubscribe linux-block" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
