Fajar A. Nugraha schrieb: > On Sat, Mar 14, 2009 at 9:40 PM, Florian Manschwetus > <[email protected]> wrote: >>> I'm using pygrub with no problems at all. I use zvols directly as >>> domU's storage (without partition). >> would be like: >> /dev/dsk/c0t600A0B800049E902000007FD49B54652d0p1 >> (what works well for the domU if I map it as xvda1) >> >> normally I map >> /dev/dsk/c0t600A0B800049E902000007FD49B54652d0p0 directly as xvda to >> domU so the domU has a full disk to partition on its own. The ...p0 I use as full disk (xvda) > > It depends more on your preferences, actually. > You could map a file (.img, .vmdk, whatever), a zvol, a disk, or a > partition as domU's disk or partition. So mapping > /dev/dsk/c0t600A0B800049E902000007FD49B54652d0p0 as xvda1 should work > as well. But that would somewhat limit scalability since xvda1 will be > limited to that size (i.e. you can't grow it), which is why I prefer > to use zvol (on opensolaris) or LVM (on Linux). Those device are mapped from FC SAN. > > pygrub can work with block devices mapped as disk (xvda) or partition > (xvda1), so I'm curious why you're experiencing problem. > so why he always say that he is unable to find anything (what is about symlinks for the kernel?)
>> I have a /boot/grub/grub.conf > > Try symlink it to menu.lst. Should work. Those symlink is already in place, but I'll alter it to use a base path, not relative one. > >> I have no initrd (isn't needed normally) > > So I take it you're not using RHEL/Centos or Debian/Ubuntu domU, or > you're building your own kernel? By default those systems need initrd > to boot correctly. > Yes I build my kernel on my own it is normal for gentoo linux >>From pygrub's perspective it doesn't really matter though, it can work > with or without initrd. > So where is the problem? I pygrub unable to read the ext3 partition when the dom0 is solaris nevada (SXCE b109)? > Regards, > > Fajar > Regards, Florian
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ xen-discuss mailing list [email protected]
