On Mon, 2008-09-29 at 16:56 -0300, Peter Cordes wrote: > tags 500593 patch > thanks > > I fixed this by re-adding the imflag ? imaxpct : defaut logic. > I also re-arranged the -N printing logic to use the values stored in sbp, > instead of duplicating code when it's easy to avoid. see attached patch. > I'm not sure you'll like it, since only some of the printed values are > re-extracted from sbp. I left ones that weren't trivial as they are. > > I also moved the code that sets dsunit and dswidth from xlv_dsunit > and xlv_dswidth to before the call to calc_default_ag_geometry(), so > multidisk will be true when the user specifies RAID parameters, not > just when they're read from LVM (xlv_ds...). Is this sensible, or is > agcount=4 appropriate for an 8 disk RAID6 on a quad-core machine? I > was thinking agcount=8 might be better, but that agcount=32 might > still be too many. I'm not expecting a huge amount of extremely small > files, and I am planning to mount with inode64...
Thanks Peter - I've fwd'd on your patch to the SGI folk and will update xfsprogs once they merge it -- thanks for the fix! -- Nathan -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
