On Mar 18,  2:20pm, John Nemeth wrote:
} On Mar 18, 12:40pm, Ronald Roskens wrote:
} } On Fri, 2007-10-26 at 14:14 +0300, Jukka Marin wrote:
} } > 
} } > I will try setting up a new server with a 3 TB (or so) RAID disk to be used
} } > for XEN virtual machines.  What is the best way of partitioning the disk -
} } > is it possible to disklabel disks this big in the normal way or should I be
} } > using wedges or something?  Do wedges work with XEN?
} } 
} } While I'm still new to setting up Xen under netbsd, here's some things
} } I've learned:
} } 
} } 2- The standard XEN3_DOM{0,U} kernels do not include the options for
} } DKWEDGE_* and drvctl.
} } 
} } wedges seem to be pretty cutting edge, and theres not a lot of
} } documentation written about them. In particular, the ptype argument for
} } dkctl addwedge isnt well defined. The types arent the same as their
} } disklabel versions. (ie: ffs = 4.2BSD, lfs = 4.4LFS)
} } 
} } You need to use gpt to create a GPT on the disk, but raidframe devices
} } do not seem to be supported. If your not using raidframe, then GPT would
} } seem to give you the flexibility in terms of the number of partitions it
} } would support as wedges.
} 
}      This is because raidframe doesn't support drvctl.  It shouldn't be
} too hard to add, but I don't have time to do it at this moment.  Greg
} could probably find the correct spot to add the code a lot faster then
} I could.

     The code has now been added.  In /usr/src/sys/dev/raidframe can
you issue the following command:

cvs up -kk -j 1.233 -j 1.234 rf_netbsdkintf.c

You will probably also want

cvs up -kk -j 1.227 -j 1.228 rf_netbsdkintf.c

You will need to add 'pseudo-device     drvctl' to your kernel
configuration.  After these steps, you can build and install a new
kernel.  Then in /dev, type './MAKEDEV drvctl'.  At this point,
reboot.  At booting, type 'drvctl -p raid0'.  You should get some XML
output describing the geometry of raid0.  If this all works, then go
ahead and try gpt.  Let us know how it works out.  Thanks.

}-- End of excerpt from John Nemeth