carlopmart wrote:
Tom Brown wrote:
Rudi Ahlers wrote:
Hi all
This is the first time I try and get Xen working, and want to install a few CentOS 5.1 32bit and FreeBSD 6.1 VM's on my CentOS 5.1 x64 server.
Trying to get the first one to install is rather trivial, and I keep on running into errors. At fist I wanted to install it fro the cdrom, so I used "virt-install --cdrom=/dev/hdc" It booted up, and I could choose my language, but then got an error saying I don't have the right driver for the cdrom install. Odd.
So, I copied the whole CD to my intranet server, to try a HTTP installation, which got much further, but fails as soon as it starts copying the files to the xvda partition.
Here's what I've done: virt-install -p --location=http://intranet/kickstart/centos/5.1/32/os/i386 --ram=512 --file=/vm/vm03.img --name=vm03
Probing for video card: Unable to probe No video hardware found, assuming headless Graphical installation not available... Starting text mode. Welcome to CentOS
+------------------+ Package Installation +------------------+ | | | Name : libXt-1.0.2-3.1.fc6-i386 | | Size : 359k | | Summary: X.Org X11 libXt runtime library | | | | | | Status: | | | | 53% | | | | Packages Bytes Time | | Total : 349 564M 0:02:30 | | Completed: 123 200M 0:00:53 | | Remaining: 226 364M 0:01:37 | | | | 35%
Guest installation complete... restarting guest. libvir: Xen Daemon error : POST operation failed: (xend.err "Error creating domain: Boot loader didn't return any data!") exception was: virDomainCreate() failed POST operation failed: (xend.err "Error creating domain: Boot loader didn't return any data!") Domain installation may not have been successful. If it was, you can restart your domain by running 'virsh start vm03'; otherwise, pleasels | <F12> next screen restart your installation. [root@gimbli ~]# virsh start vm03 libvir: Xen Daemon error : POST operation failed: (xend.err "Error creating domain: Boot loader didn't return any data!") error: Failed to start domain vm03
/var/log/messages got filled with the following messages:
stem_r:xenstored_t:s0 tcontext=system_u:system_r:xenstored_t:s0 tclass=capability Feb 29 10:36:59 gimbli kernel: audit(1204274218.949:2139): avc: denied { sys_resource } for pid=2445 comm="xenstored" capability=24 scontext=system_u:system_r:xenstored_t:s0 tcontext=system_u:system_r:xenstored_t:s0 tclass=capability Feb 29 10:36:59 gimbli kernel: audit(1204274218.949:2140): avc: denied { sys_resource } for pid=2445 comm="xenstored" capability=24 scontext=system_u:system_r:xenstored_t:s0 tcontext=system_u:system_r:xenstored_t:s0 tclass=capability Feb 29 10:36:59 gimbli kernel: audit(1204274218.949:2141): avc: denied { sys_resource } for pid=2445 comm="xenstored" capability=24 scontext=system_u:system_r:xenstored_t:s0 tcontext=system_u:system_r:xenstored_t:s0 tclass=capability Feb 29 10:36:59 gimbli kernel: audit(1204274218.949:2142): avc: denied { sys_resource } for pid=2445 comm="xenstored" capability=24 scontext=system_u:system_r:xenstored_t:s0 tcontext=system_u:system_r:xenstored_t:s0 tclass=capability
What am I doing wrong?
as far as i am aware on standard RHEL or CentOS 5 packages a 32bit domU on a 64bit dom0 is not supported and known to crash
That's incorrent. I have three Rhel5.1 x86_64 xenhosts with different 32 bits domU, centos 5.1, centos 5.0 and rhel5.1 included ...
I think that problem is selinux. In redhat docs explains that domU needs to reside on /var/lib/xen if you use selinux. You can try to disable selinux (/etc/sysconfig/selinux) and all works ok.
I have got selinux disabled.