On 17/12/2009, at 11:02 PM, Henning Brauer wrote: > * Aaron Mason <simplersolut...@gmail.com> [2009-12-17 03:50]: >> From what I've seen, 4k blocks are supported by most filesystems >> anyway - and besides, provided the partitions are created on 4k block >> boundaries, there's no reason for any concern IMHO. > > you are missing the point. these drives do not expose the fact that > they use 4K blocks internally, they pretend to work with the > traditional 512byte blocks. And there is a trap there, as the original > article points out.
this is not a new problem. RAID controllers have been managing io in chunks larger than 512bytes for years while still presenting 512byte blocks to the machines that use them. they definitely behave better if you do can match your io sizes to the large chunk size they use behind the scenes, but they still work fine if you want to write a random 512 byte block anywhere on them. it is true openbsd could make some effort to figure out what block size a disk would like you to use when talking to it, but at the same time it isnt like the machine isnt going to boot if we dont bother. let's not wave our hands around in the air for no good reason. surely there are more important things to get upset about. dlg
