Cannot start domU from NFS

Cannot start domU from NFS

[Nauzad Sadry]

Hello everyone
I have Xen-2.0.1 setup with dom0 starting from local disk.
I need to start domU from NFS
The NFS server works - both local and remote clients can use it.
But not domU 

Here is my config
======= /etc/xen/vm_config_nfs =========
kernel = "/boot/vmlinuz-2.6.9-xenU"
ramdisk = "/boot/initrd-fc3.img"
memory = 128
name = "VM-1"
nics=3
dhcp="dhcp"
root = "/dev/nfs"
nfs_server = "192.168.1.123"
nfs_root   = '/xen_nfs/root'
extra = "4 enforcing=0"
======================================

domU starts booting and hangs at mounting its root 

Here is the boot log

Using config file "/etc/xen/vm_config_nfs".
Started domain VM-1, console on port 9608
************ REMOTE CONSOLE: CTRL-] TO QUIT ********
Linux version 2.6.9-xenU (xenod@labyrinth.cl.cam.ac.uk) (
gcc version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)) #1 We
d Nov 17 22:29:37 GMT 2004
BIOS-provided physical RAM map:
 Xen: 0000000000000000 - 0000000008000000 (usable)
128MB LOWMEM available.
DMI not present.
Built 1 zonelists
Kernel command line:  ip=:192.168.123.123:::VM-1:eth0:dhcp
root=/dev/nfs nfsroot=192.168.123.123:/xen_nfs/root 3 VMID=1
enforcing=0
Initializing CPU#0
PID hash table entries: 1024 (order: 10, 16384 bytes)
Xen reported: 1793.416 MHz processor.
Using tsc for high-res timesource
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 126148k/131072k available (1596k kernel code, 4860k reserved,
466k data, 92k init, 0k highmem)
Checking if this processor honours the WP bit even in supervisor mode... Ok.
Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
CPU: L2 Cache: 1024K (64 bytes/line)
CPU: AMD Opteron(tm) Processor 244 stepping 08
Enabling unmasked SIMD FPU exception support... done.
Checking 'hlt' instruction... disabled
checking if image is initramfs... it is
Freeing initrd memory: 1061k freed
NET: Registered protocol family 16
Initializing Cryptographic API
RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
Xen virtual console successfully installed as tty
Event-channel device installed.
Starting Xen Balloon driver
xen_blk: Initialising virtual block device driver
xen_blk: Timeout connecting to device!
xen_net: Initialising virtual ethernet driver.
NET: Registered protocol family 2
IP: routing cache hash table of 1024 buckets, 8Kbytes
TCP: Hash tables configured (established 8192 bind 16384)
NET: Registered protocol family 1
NET: Registered protocol family 17
IP-Config: Incomplete network configuration information.
Freeing unused kernel memory: 92k freed
Red Hat nash version 4.1.18 starting
Mounted /proc filesystem
Mounting sysfs
Creating /dev
Starting udev
Creating root device
Mounting root filesystem
mount: error 6 mounting ext3
mount: error 2 mounting none
Switching to new root
switchroot: mount failed: 22
umount /initrd/dev failed: 2
Kernel panic - not syncing: Attempted to kill init!
 <0>Rebooting in 1 seconds..Linux version 2.6.9-xenU (xen


On my NFS export, the etc/fstab is as follows
/xen_nfs/root        /                           ext3    defaults        1 1
none                    /dev/pts                devpts  gid=5,mode=620  0 0
none                    /proc                    proc    defaults        0 0
none                    /dev/shm              tmpfs   defaults        0 0
/dev/hda5             swap                    swap    defaults        0 0
/dev/cdrom           /mnt/cdrom           udf,iso9660 noauto,owner,kudzu,ro 0 0
/dev/fd0                /mnt/floppy           auto    noauto,owner,kudzu 0 0

Help will be really appreciated 

Thanks

Nauzad


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: Cannot start domU from NFS

[Christian Limpach]

On Thu, Jan 06, 2005 at 11:37:09PM -0800, Nauzad Sadry wrote:
> Hello everyone
> I have Xen-2.0.1 setup with dom0 starting from local disk.
> I need to start domU from NFS
> The NFS server works - both local and remote clients can use it.
> But not domU 
> 
> Here is the boot log
> 
> IP-Config: Incomplete network configuration information.

Try booting with a static IP configuration.  For dhcp you will need to
build your own xenU kernel since the one we provide doesn't have dhcp
client support compiled in.  Alternatively you could use our xen0 kernel
which has dhcp client support compiled in.

> On my NFS export, the etc/fstab is as follows
> /xen_nfs/root        /                           ext3    defaults        1 1
> none                    /dev/pts                devpts  gid=5,mode=620  0 0
> none                    /proc                    proc    defaults        0 0
> none                    /dev/shm              tmpfs   defaults        0 0
> /dev/hda5             swap                    swap    defaults        0 0
> /dev/cdrom           /mnt/cdrom           udf,iso9660 noauto,owner,kudzu,ro 0 0
> /dev/fd0                /mnt/floppy           auto    noauto,owner,kudzu 0 0

Unless you're exporting a block device to the domain as hda5, this will
fail when swap gets enabled.  Also, your entry for the root fs is not
correct.  The 1st field should be of the form server:/path/to/root -- 
I usually use IP addresses, a hostname might work, certainly if it's
defined in /etc/hosts.  The filesystem type needs to be nfs, not ext3.
Also the dump frequency and fsck pass number fields should both be 0.

    christian



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: Cannot start domU from NFS

[Nauzad Sadry]

Hi Christian

Thanks for the reply, but your sugesstions did not work .. i am still
failing at the same point. It is somehow not able to mount the root
file system.

1. Do I need to have "/dev/nfs" as an existing device in the target
(domain 0)system
2. Are there any daemons or services that I need running on my domain0
before attempting to create domain1
3. When you say use static IP configuration, I presume this means setting 
    ip="<address> in the xen config file.
4. After completing IP config I believe it should use contact the NFS
bootserver, but i do not see any such commands in the boot log.

Please help me if something is wrong in my configuration

FYR, here is the boot log

************ REMOTE CONSOLE: CTRL-] TO QUIT ********
Linux version 2.6.9-xenU (xenod@labyrinth.cl.cam.ac.uk) (gcc
version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)) #1 Wed Nov 17
 22:29:37 GMT 2004
BIOS-provided physical RAM map:
 Xen: 0000000000000000 - 0000000008000000 (usable)
128MB LOWMEM available.
DMI not present.
Built 1 zonelists
Kernel command line:  ip=192.168.123.145:192.168.123.200:192.
168.123.254:255.255.255.0:VM-3:eth0:off root=/dev/nfs nfsroot
=192.168.123.200:/xen_nfs/tmp 3 VMID=3 enforcing=0
Initializing CPU#0
PID hash table entries: 1024 (order: 10, 16384 bytes)
Xen reported: 1793.261 MHz processor.
Using tsc for high-res timesource
Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
Memory: 126148k/131072k available (1596k kernel code, 4860k reserved,
466k data, 92k init, 0k highmem)
Checking if this processor honours the WP bit even in supervisor mode... Ok.
Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
CPU: L2 Cache: 1024K (64 bytes/line)
CPU: AMD Opteron(tm) Processor 244 stepping 08
Enabling unmasked SIMD FPU exception support... done.
Checking 'hlt' instruction... disabled
checking if image is initramfs... it is
Freeing initrd memory: 1061k freed
NET: Registered protocol family 16
Initializing Cryptographic API
RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
Xen virtual console successfully installed as tty
Event-channel device installed.
Starting Xen Balloon driver
xen_blk: Initialising virtual block device driver
xen_blk: Timeout connecting to device!
xen_net: Initialising virtual ethernet driver.
NET: Registered protocol family 2
IP: routing cache hash table of 1024 buckets, 8Kbytes
TCP: Hash tables configured (established 8192 bind 16384)
NET: Registered protocol family 1
NET: Registered protocol family 17
IP-Config: Complete:
      device=eth0, addr=192.168.123.145, mask=255.255.255.0,
gw=192.168.123.254, host=VM-3, domain=, nis-domain=(none),
     bootserver=192.168.123.200, rootserver=192.168.123.200,
rootpath=
Freeing unused kernel memory: 92k freed
Red Hat nash version 4.1.18 starting
Mounted /proc filesystem
Mounting sysfs
Creating /dev
Starting udev
Creating root device
Mounting root filesystem
mount: error 6 mounting ext3
mount: error 2 mounting none
Switching to new root
switchroot: mount failed: 22
umount /initrd/dev failed: 2
Kernel panic - not syncing: Attempted to kill init!



On Fri, 7 Jan 2005 10:01:19 +0000, Christian Limpach
<Christian.Limpach@cl.cam.ac.uk> wrote:
> On Thu, Jan 06, 2005 at 11:37:09PM -0800, Nauzad Sadry wrote:
> > Hello everyone
> > I have Xen-2.0.1 setup with dom0 starting from local disk.
> > I need to start domU from NFS
> > The NFS server works - both local and remote clients can use it.
> > But not domU
> > 
> > Here is the boot log
> >
> > IP-Config: Incomplete network configuration information.
> 
> Try booting with a static IP configuration.  For dhcp you will need to
> build your own xenU kernel since the one we provide doesn't have dhcp
> client support compiled in.  Alternatively you could use our xen0 kernel
> which has dhcp client support compiled in.
> 
> > On my NFS export, the etc/fstab is as follows
> > /xen_nfs/root        /                           ext3    defaults        1 1
> > none                    /dev/pts                devpts  gid=5,mode=620  0 0
> > none                    /proc                    proc    defaults        0 0
> > none                    /dev/shm              tmpfs   defaults        0 0
> > /dev/hda5             swap                    swap    defaults        0 0
> > /dev/cdrom           /mnt/cdrom           udf,iso9660 noauto,owner,kudzu,ro 0 0
> > /dev/fd0                /mnt/floppy           auto    noauto,owner,kudzu 0 0
> 
> Unless you're exporting a block device to the domain as hda5, this will
> fail when swap gets enabled.  Also, your entry for the root fs is not
> correct.  The 1st field should be of the form server:/path/to/root --
> I usually use IP addresses, a hostname might work, certainly if it's
> defined in /etc/hosts.  The filesystem type needs to be nfs, not ext3.
> Also the dump frequency and fsck pass number fields should both be 0.
> 
>    christian
> 
>


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: Is continuous replication of state possible?

[Harry Butterworth]

Ian Pratt wrote:
> Software-implemented hardware fault-tolerance is on the Xen
> research roadmap.
>
> It basically just requires deterministic execution and event
> injection. Doing this for uniprocessor guests is fairly straight
> forward. Doing it for SMP guests (with decent performance) is
> going to be a huge challenge, as determinism is hard to achieve. We're
> looking in to it...

I thought a little about trying to implement a byzantine fault tolerant
SMP virtual machine using state replication.  I read Barbara Liskov's
group's work for some BFT distributed consensus protocols and there is
some good work out there on BFT routing in ad-hoc networks and a little
bit on BFT authentication but I didn't find much on the SMP issues.

I wanted to try out two approaches:

1) Putting each CPU into a separate BFT context and suffering the
consensus protocol overhead in inter-cpu communications.
I thought this would probably work OK for replicated servers which were
relatively near each other (for example on IBM's proposed CIB hardware)
but I was also interested in global replication for continuity through
disasters so I was considering
2) Assuming a standard execution rate for instructions and speculatively
executing them non-deterministically in parallel. Non-determinism is
only significant when communication between CPUs happens in the wrong
order (for example when an instruction speculatively assigned to a
particular time-step on one cpu writes to a cache line which has already
been read from by an instruction speculatively assigned to a later
time-step by another cpu) this must be detected at which point the local
replica has to roll-back and then roll forwards more carefully to get
the serialisation right. I think this would work better when the
replicas are separated geographically because there is no BFT protocol
overhead for inter-cpu communication.

In both cases, I was going to emulate CPUs rather than try to use the
real CPU or any hardware features.  With emulation, I wanted to try out
using a high emulated:real CPU ratio and/or deep pipelines to ameliorate
the inherent stall problem of the first approach.  Emulation seems to be
a requirement for investigating the second approach.

I think there's also a possibility of using dynamic translation to make
either of these approaches go faster. With dynamic translation, either
approach might get to within a small factor of the theoretical limit for
BFT.

Not good for scientific performance computing but the BFT, global
replication, massive parallelism and single system image features would
be great for a lot of business applications.

I'm very curious, do you have another approach? Any thoughts about
real-time?

Harry.


-- 
Harry Butterworth <harry@hebutterworth.freeserve.co.uk>



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt

Re: Cannot start domU from NFS

[Nauzad Sadry]

Hello guys

I was able to sucessfully start domU using NFS. I had to disable the
ramdisk (=initrd-fc3.img) in the vm configuration file. The ramdisk
was preventing domain0 from using NFS to mounting the root

Thanks for your help

Nauzad

On Sat, 8 Jan 2005 16:42:21 -0800, Nauzad Sadry <nauzad@gmail.com> wrote:
> Hi Christian
> 
> Thanks for the reply, but your sugesstions did not work .. i am still
> failing at the same point. It is somehow not able to mount the root
> file system.
> 
> 1. Do I need to have "/dev/nfs" as an existing device in the target
> (domain 0)system
> 2. Are there any daemons or services that I need running on my domain0
> before attempting to create domain1
> 3. When you say use static IP configuration, I presume this means setting
>    ip="<address> in the xen config file.
> 4. After completing IP config I believe it should use contact the NFS
> bootserver, but i do not see any such commands in the boot log.
> 
> Please help me if something is wrong in my configuration
> 
> FYR, here is the boot log
> 
> ************ REMOTE CONSOLE: CTRL-] TO QUIT ********
> Linux version 2.6.9-xenU (xenod@labyrinth.cl.cam.ac.uk) (gcc
> version 3.2.2 20030222 (Red Hat Linux 3.2.2-5)) #1 Wed Nov 17
> 22:29:37 GMT 2004
> BIOS-provided physical RAM map:
> Xen: 0000000000000000 - 0000000008000000 (usable)
> 128MB LOWMEM available.
> DMI not present.
> Built 1 zonelists
> Kernel command line:  ip=192.168.123.145:192.168.123.200:192.
> 168.123.254:255.255.255.0:VM-3:eth0:off root=/dev/nfs nfsroot
> =192.168.123.200:/xen_nfs/tmp 3 VMID=3 enforcing=0
> Initializing CPU#0
> PID hash table entries: 1024 (order: 10, 16384 bytes)
> Xen reported: 1793.261 MHz processor.
> Using tsc for high-res timesource
> Dentry cache hash table entries: 32768 (order: 5, 131072 bytes)
> Inode-cache hash table entries: 16384 (order: 4, 65536 bytes)
> Memory: 126148k/131072k available (1596k kernel code, 4860k reserved,
> 466k data, 92k init, 0k highmem)
> Checking if this processor honours the WP bit even in supervisor mode... Ok.
> Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
> CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
> CPU: L2 Cache: 1024K (64 bytes/line)
> CPU: AMD Opteron(tm) Processor 244 stepping 08
> Enabling unmasked SIMD FPU exception support... done.
> Checking 'hlt' instruction... disabled
> checking if image is initramfs... it is
> Freeing initrd memory: 1061k freed
> NET: Registered protocol family 16
> Initializing Cryptographic API
> RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
> Xen virtual console successfully installed as tty
> Event-channel device installed.
> Starting Xen Balloon driver
> xen_blk: Initialising virtual block device driver
> xen_blk: Timeout connecting to device!
> xen_net: Initialising virtual ethernet driver.
> NET: Registered protocol family 2
> IP: routing cache hash table of 1024 buckets, 8Kbytes
> TCP: Hash tables configured (established 8192 bind 16384)
> NET: Registered protocol family 1
> NET: Registered protocol family 17
> IP-Config: Complete:
>      device=eth0, addr=192.168.123.145, mask=255.255.255.0,
> gw=192.168.123.254, host=VM-3, domain=, nis-domain=(none),
>     bootserver=192.168.123.200, rootserver=192.168.123.200,
> rootpath=
> Freeing unused kernel memory: 92k freed
> Red Hat nash version 4.1.18 starting
> Mounted /proc filesystem
> Mounting sysfs
> Creating /dev
> Starting udev
> Creating root device
> Mounting root filesystem
> mount: error 6 mounting ext3
> mount: error 2 mounting none
> Switching to new root
> switchroot: mount failed: 22
> umount /initrd/dev failed: 2
> Kernel panic - not syncing: Attempted to kill init!
> 
> On Fri, 7 Jan 2005 10:01:19 +0000, Christian Limpach
> <Christian.Limpach@cl.cam.ac.uk> wrote:
> > On Thu, Jan 06, 2005 at 11:37:09PM -0800, Nauzad Sadry wrote:
> > > Hello everyone
> > > I have Xen-2.0.1 setup with dom0 starting from local disk.
> > > I need to start domU from NFS
> > > The NFS server works - both local and remote clients can use it.
> > > But not domU
> > >
> > > Here is the boot log
> > >
> > > IP-Config: Incomplete network configuration information.
> >
> > Try booting with a static IP configuration.  For dhcp you will need to
> > build your own xenU kernel since the one we provide doesn't have dhcp
> > client support compiled in.  Alternatively you could use our xen0 kernel
> > which has dhcp client support compiled in.
> >
> > > On my NFS export, the etc/fstab is as follows
> > > /xen_nfs/root        /                           ext3    defaults        1 1
> > > none                    /dev/pts                devpts  gid=5,mode=620  0 0
> > > none                    /proc                    proc    defaults        0 0
> > > none                    /dev/shm              tmpfs   defaults        0 0
> > > /dev/hda5             swap                    swap    defaults        0 0
> > > /dev/cdrom           /mnt/cdrom           udf,iso9660 noauto,owner,kudzu,ro 0 0
> > > /dev/fd0                /mnt/floppy           auto    noauto,owner,kudzu 0 0
> >
> > Unless you're exporting a block device to the domain as hda5, this will
> > fail when swap gets enabled.  Also, your entry for the root fs is not
> > correct.  The 1st field should be of the form server:/path/to/root --
> > I usually use IP addresses, a hostname might work, certainly if it's
> > defined in /etc/hosts.  The filesystem type needs to be nfs, not ext3.
> > Also the dump frequency and fsck pass number fields should both be 0.
> >
> >    christian
> >
> >
>


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt