Matthias Schniedermeyer wrote:
My first guess would be "aggregategroups".
Just to be pedantic, "AG" is "Allocation Group" :)
For me mkfs.xfs ALWAYS uses a too low agcount, with a too low agcount a
single aggregategroup can be lager than 4GB which i was told is a "no no".
That should be fine now (as of xfsprogs 2.6.0)
mkfs_xfs.c: revision 1.52
date: 2003/10/28 04:41:37; author: nathans; state: Exp; lines: +141 -126
modid: xfs-cmds:slinx:160712a
Rework the mkfs allocation group sizing algorithm, making better use of
the available bits. This changes the maximum allocation group size
enforced by mkfs to be 1TB (from 4GB), which scales alot better for very
large filesystems.
Hm, Nathan, I think we need to update the mkfs man page...?
Do a xfs_info on your device and if (agsize * sectsz) > 4GB (or
"Capacity / agcount" > 4GB) then you have to reformat with a bigger
agcount/lower agsize.
Feel free to try, maybe it will help, but the design intent is that
larger AGs should be fine.
-Eric
|