[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-users] Anyone able to NFS boot on xen 4.x ?


  • From: Bruce Edge <bruce.edge@xxxxxxxxx>
  • Date: Wed, 9 Jun 2010 16:17:59 -0700
  • Cc: xen-users@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Wed, 09 Jun 2010 16:19:53 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:cc :content-type; b=w+ByKsCVjOJhLs3++Mn2oejA0DNSVd12+97zvqd5YciU4I8Pf20Job1GQIOp3nhHa/ rZAGVJkEwTvbesQYN1w9Giq1z7MlXceMTItVr8yMFqTFq2Q4jLdRCyDnIBx6KjHpn2xh HKnceGd8G8vgrUzCj+mjNL/pC7gn7wsMZ08BY=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

I converted to the pv-ops kernel and got a lot further, but the domU NFS boot is still hanging up here:

[    2.510039] IP-Config: Got DHCP answer from 135.149.78.198, my address is 135.149.77.123
[    2.510604] IP-Config: Complete:
[    2.510611]      device=eth0, addr=135.149.77.123, mask=255.255.252.0, gw=135.149.76.1,
.....
[    2.510763] Looking up port of RPC 100003/2 on 135.149.75.134
[    5.510807] Looking up port of RPC 100005/1 on 135.149.75.134

- hangs here, or takes a long to to progress to:

[   14.670416] Looking up port of RPC 100005/1 on 135.149.75.134
[  222.440118] VFS: Mounted root (nfs filesystem) on device 0:14.
[  222.440332] Freeing unused kernel memory: 712k freed
[  222.440553] Write protecting the kernel read-only data: 8448k

- then hangs.

Here's the config summary:
        root       = /dev/nfs
        nfs_server = 135.149.75.134
        nfs_root   = /home/nfsroot/root,tcp
        initrd     = /boot/initrd.img-2.6.32.14.nfs
        mode         = nfs
        boot         = n
        device_model = /usr/lib/xen/bin/qemu-dm
        builder      = linux
        kernel       = /boot/vmlinuz-2.6.32.14
        memory       = 4096
        extra        =   kdb=on nmi_watchdog=1 init_cpus=0 console=ttyS0 root=/dev/nfs ip=dhcp
        pci          = ['10:00.3@10', '10:00.2@11', '10:00.1@12', '10:00.0@13', '07:00.3@14', '07:00.2@15', '07:00.1@16', '07:00.0@17', '17:00.0@18', '17:00.1@19', '14:00.0@1a', '14:00.1@1b', '0a:00.0@1c', '0a:00.1@1d']
        vif          = ['model=e1000, bridge=eth0, mac=00:16:3e:00:22:01,vifname=vif-dpm']


-Bruce


On Sat, May 29, 2010 at 2:19 AM, Felix Kuperjans <felix@xxxxxxxxxxxxxxxxxx> wrote:
Right, I am using pv_ops domU as a pv kernel.

I never actually booted a hvm with NFS root, but it should be possible if you boot it like a native NFS root system, i.e. network boot with NFS root options to the kernel.
Those options in the domU config file will most likely have no effect, you need to setup a network boot server with NFS root and then boot the hvm via network (this is supported by qemu-dm).

You will find some tutorials on setting up native machines for root filesystem over NFS and those should be applicable to XEN HVM, too.

I hope this helps,
Felix

Am 29.05.2010 01:44, schrieb Bruce Edge:
> Felix,
> Thanks for the info.
>
> I assume form the
>      > kernel     = "/mnt/public/xen/linux-rescue"
> line that you are using a pv domU.
>
> Can one NFS boot an hvm system?
>
> -Bruce
>
>
>
> On Fri, May 28, 2010 at 10:13 AM, Felix Kuperjans
> <felix@xxxxxxxxxxxxxxxxxx> wrote:
>
>> I am using NFS root filesystem for many of my domUs on XEN 4.0.0 for
>> rescue systems.
>>
>> It's just as it should be:
>>
>> domu.cfg:
>>
>> # General information:
>> name       = "Rescue"
>> memory     = 2048
>> vcpus      = 4
>>
>> # Boot configuration:
>> kernel     = "/mnt/public/xen/linux-rescue"
>> root       = "/dev/nfs"
>> nfs_root   = "/mnt/rescue"
>>
>> # Additional configuration:
>> nographic  = "yes"
>> timer_mode = 1
>>
>> # Kernel network configuration:
>> interface  = "eth0"
>> ip         = "10.157.0.1"
>> netmask    = "255.252.0.0"
>> gateway    = "10.156.0.1"
>> hostname   = "rescue.domain"
>> nfs_server = "10.156.0.1"
>>
>> # Network interfaces:
>> vif        = [ "mac=00:16:3e:9d:00:01,vifname=vif-rescue" ]
>>
>> # Virtual disks:
>> disk       = [ ]
>>
>>
>> Some things you need to do:
>>
>> 1. Root filesystem over NFS support in the domU kernel
>> 2. NFS export on Dom0 ro or rw and only IP-based authentification
>> 3. Use NFSv3; NFSv4 has no root-fs support without an initrd and many
>> extra settings.
>> 4. Network settings must be correct at boot time (kernel level
>> autoconfiguration), so have to adopt my settings to your network setup
>> (mine is routing with internal IPs, NAT and much firewalling, far away
>> from default bridging)
>>
>> Regards,
>> Felix Kuperjans
>>
>> Am 28.05.2010 18:50, schrieb Bruce Edge:
>>
>>> Has anyone been able to get domU NFS boots working with any version of
>>> Xen 4.x? If so, can you please post your config? Both the dom0 Xen
>>> version & kernel, as well as the domU config file.
>>>
>>> I've spent a lot of time running through all the docs, HOWTO's and
>>> published configs and assorted patches for Ubuntu 9x-10x and Xen 4x
>>> trying NFS booting with each. I'm not going to document all the
>>> failure modes here as I've already posted most to the xen-devel list.
>>>
>>> I'm starting to think that it plain doesn't work with Xen 4. Does
>>> anyone else use this?
>>>
>>> Thanks in advance.
>>>
>>> -Bruce
>>>
>>> _______________________________________________
>>> Xen-users mailing list
>>> Xen-users@xxxxxxxxxxxxxxxxxxx
>>> http://lists.xensource.com/xen-users
>>>
>>>
>>>
>> _______________________________________________
>> Xen-users mailing list
>> Xen-users@xxxxxxxxxxxxxxxxxxx
>> http://lists.xensource.com/xen-users
>>
>>
> _______________________________________________
> Xen-users mailing list
> Xen-users@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-users
>
>

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.