On Fri, 11 Apr 2003 at 6:24am, Robert Sander wrote
> When creating an XFS volume on an external hardware RAID which has a
> stripe size of 128 1024k blocks, what are the preffered values for sunit
> and swidth.
> Or is there no performance degrade when using mkfs.xfs without
> specifying sunit and swidth explicitely? In that case sunit and swidth
> are set to 0.
I'll quote below the response I got from Steve Lord way back when I asked
this. The first level of quoting is me:
\begin{quote}
> The unit's default is a stripe size of 64 512 byte blocks. The stripe
> size can be set between 4 and 128 (incrementally, of course). For the
> defaults, would this be correct:
>
> mkfs.xfs -d sunit=1,swidth=64
>
> Or should I multiply both of those numbers by 512? Also, should I make
> the logfile any bigger than the default due to the size of the partition?
>
Actually this is incorrect, you need
-d sunit=64,swidth=448
The sunit is the amount of data on each disk in 512 byte chunks, the swidth
is the total amount of data before we go back to the first disk again. You
have 8 disks in raid5 so 7 data disks and a parity disk. You need 7 * 64
as the stripe width.
\end{quote}
--
Joshua Baker-LePain
Department of Biomedical Engineering
Duke University
|