headers
Steven Hazel | 3 Jul 2008 20:40
Gravatar

[uml-user] running UML in EC2

Hi,

I've been trying to get user-mode-linux running a reasonably recent linux setup in Amazon's EC2.  The host is required to be Amazon's 2.6.16-xenU kernel.  In my case, it's running Ubuntu Gutsy.  The guest is the version of UML that comes with Ubuntu, 2.6.22-rc5.

I think the problem must be with the libc.  If I use an older libc on both the host and the guest, I can get it running.  With a newer libc on the guest, it locks up after mounting the root fs.  With a newer libc on the host, I get the error pasted below.  Any help would be appreciated!

Core dump limits :
        soft - 0
        hard - NONE
Checking that ptrace can change system call numbers...OK
Checking syscall emulation patch for ptrace...OK
Checking advanced syscall emulation patch for ptrace...OK
Checking for tmpfs mount on /dev/shm...OK
Checking PROT_EXEC mmap in /dev/shm/...OK
Checking for the skas3 patch in the host:
  - /proc/mm...not found: No such file or directory
  - PTRACE_FAULTINFO...not found
  - PTRACE_LDT...not found
UML running in SKAS0 mode
[42949372.960000] Linux version 2.6.22-rc5 (root <at> vernadsky) (gcc version 4.1.3 20070629 (prerelease) (Ubuntu 4.1.2-13ubuntu2)) #2 Mon Jul 2 10:14:22 GMT 2007
[42949372.960000] Built 1 zonelists.  Total pages: 8128
[42949372.960000] Kernel command line: ubd0=laforge-root.img root=98:0
[42949372.960000] PID hash table entries: 128 (order: 7, 512 bytes)
[42949372.960000] Dentry cache hash table entries: 4096 (order: 2, 16384 bytes)
[42949372.960000] Inode-cache hash table entries: 2048 (order: 1, 8192 bytes)
[42949372.960000] Memory: 29312k available
[42949374.140000] Mount-cache hash table entries: 512
[42949374.140000] Checking for host processor cmov support...Yes
[42949374.140000] Checking for host processor xmm support...No
[42949374.140000] Checking that host ptys support output SIGIO...Yes
[42949374.140000] Checking that host ptys support SIGIO on close...No, enabling workaround
[42949374.140000] Using 2.6 host AIO
[42949374.140000] NET: Registered protocol family 16
[42949374.140000] NET: Registered protocol family 2
[42949374.430000] IP route cache hash table entries: 1024 (order: 0, 4096 bytes)
[42949374.430000] TCP established hash table entries: 1024 (order: 1, 8192 bytes)
[42949374.430000] TCP bind hash table entries: 1024 (order: 0, 4096 bytes)
[42949374.430000] TCP: Hash tables configured (established 1024 bind 1024)
[42949374.430000] TCP reno registered
[42949374.490000] Checking host MADV_REMOVE support...OK
[42949374.490000] mconsole (version 2) initialized on /root/.uml/mHLBJP/mconsole
[42949374.490000] Mapper v0.1
[42949374.490000] mmapper_init - find_iomem failed
[42949374.490000] UML Watchdog Timer
[42949374.490000] Host TLS support detected
[42949374.490000] Detected host type: i386
[42949374.490000] VFS: Disk quotas dquot_6.5.1
[42949374.490000] Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
[42949374.490000] JFS: nTxBlock = 229, nTxLock = 1832
[42949374.490000] SGI XFS with ACLs, security attributes, realtime, large block numbers, no debug enabled
[42949374.490000] SGI XFS Quota Management subsystem
[42949374.490000] io scheduler noop registered
[42949374.490000] io scheduler anticipatory registered (default)
[42949374.490000] io scheduler deadline registered
[42949374.490000] io scheduler cfq registered
[42949374.650000] SoftDog: cannot register miscdev on minor=130 (err=-16)
[42949374.650000] TCP cubic registered
[42949374.650000] NET: Registered protocol family 1
[42949374.650000] NET: Registered protocol family 17
[42949374.650000] Initialized stdio console driver
[42949374.650000] Console initialized on /dev/tty0
[42949374.650000] Initializing software serial port version 1
[42949374.710000]  ubda: unknown partition table
[42949374.750000] VFS: Mounted root (ext2 filesystem) readonly.
[42949374.950000] Kernel panic - not syncing: start_userspace : expected SIGSTOP, got status = 2943
[42949374.950000]
[42949374.950000] EIP: 0073:[<bfffe402>] CPU: 0 Not tainted ESP: 007b:b7de4fc0 EFLAGS: 00200246
[42949374.950000]     Not tainted
[42949374.950000] EAX: 00000000 EBX: 00000ab2 ECX: 00000013 EDX: 00000ab2
[42949374.950000] ESI: 00000aae EDI: 00000011 EBP: b7de4fd8 DS: 007b ES: 007b
[42949374.950000] 08327e0c:  [<0806ce8c>] show_regs+0xb4/0xb9
[42949374.950000] 08327e38:  [<0805b038>] panic_exit+0x25/0x3f
[42949374.950000] 08327e4c:  [<0807c418>] notifier_call_chain+0x21/0x46
[42949374.950000] 08327e6c:  [<0807c4b3>] __atomic_notifier_call_chain+0x17/0x19
[42949374.950000] 08327e88:  [<0807c4ca>] atomic_notifier_call_chain+0x15/0x17
[42949374.950000] 08327ea4:  [<08071532>] panic+0x52/0xd8
[42949374.950000] 08327ec4:  [<0806a3f3>] start_userspace+0xd0/0x11c
[42949374.950000] 08327ee4:  [<0805b32a>] init_new_context_skas+0x107/0x14a
[42949374.950000] 08327f04:  [<080ae8da>] do_execve+0x89/0x1cb
[42949374.950000] 08327f24:  [<08057ced>] execve1+0x2a/0x54
[42949374.950000] 08327f48:  [<08057d2c>] um_execve+0x15/0x41
[42949374.950000] 08327f64:  [<08059e25>] kernel_execve+0x2c/0x37
[42949374.950000] 08327f84:  [<080576cf>] run_init_process+0x1b/0x20
[42949374.950000] 08327f98:  [<08057764>] init_post+0x90/0xb8
[42949374.950000] 08327fa4:  [<08049879>] kernel_init+0x7e/0x85
[42949374.950000] 08327fb4:  [<08067315>] run_kernel_thread+0x37/0x42
[42949374.950000] 08327fe0:  [<0805b49d>] new_thread_handler+0x57/0x7e
[42949374.950000] 08327ffc:  [<a55a5a5a>] 0xa55a5a5a
[42949374.950000]
Segmentation fault


--
Steven Hazel
sah <at> awesame.org 
-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
_______________________________________________
User-mode-linux-user mailing list
User-mode-linux-user <at> lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/user-mode-linux-user
Permalink | Reply | 0 comments |
headers
Will Stockdell | 3 Jul 2008 20:46
Picon
Favicon

[uml-user] increasing eth ids


I'm having a networking problem with a UML installation.  I can boot the
uml guest and get consoles and I can even get networking configured and
working, but when I reboot the guest it complains that it can't find the
ethernet device and /sbin/ifconfig shows that the ethernet device available
is one greater than previously.  IOW, the first time, it's eth0, the second
eth1, the third eth3, etc.  Obviously, this means
that /etc/network/interfaces doesn't have the right eth device.  I can then
manually configure the current device whether it is eth0 or eth22, but I'd
rather not have to do that every time I boot the guest.  So my question is
why does the device keep incrementing?  What controls it, and how can I
make it stop?

Thanks,

Will

------------------------------

This shirt is "dry-clean only," which means it's dirty.
                    --Mitch Hedberg

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 2 comments |
headers
Flavio | 3 Jul 2008 21:00
Picon
Gravatar

Re: [uml-user] increasing eth ids

2008/7/3 Will Stockdell <stockdel <at> us.ibm.com>:
>
> I'm having a networking problem with a UML installation.  I can boot the
> uml guest and get consoles and I can even get networking configured and
> working, but when I reboot the guest it complains that it can't find the
> ethernet device and /sbin/ifconfig shows that the ethernet device available
> is one greater than previously.  IOW, the first time, it's eth0, the second
> eth1, the third eth3, etc.  Obviously, this means
> that /etc/network/interfaces doesn't have the right eth device.  I can then
> manually configure the current device whether it is eth0 or eth22, but I'd
> rather not have to do that every time I boot the guest.  So my question is
> why does the device keep incrementing?  What controls it, and how can I
> make it stop?
Post your command line used to run UML please.

Regards,

Flavio

>
> Thanks,
>
> Will
>
>
> ------------------------------
>
> This shirt is "dry-clean only," which means it's dirty.
>                    --Mitch Hedberg
>
>
> -------------------------------------------------------------------------
> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
> Studies have shown that voting for your favorite open source project,
> along with a healthy diet, reduces your potential for chronic lameness
> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
> _______________________________________________
> User-mode-linux-user mailing list
> User-mode-linux-user <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/user-mode-linux-user

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 0 comments |
headers
Iustin Pop | 3 Jul 2008 21:12

Re: [uml-user] increasing eth ids

On Thu, Jul 03, 2008 at 01:46:44PM -0500, Will Stockdell wrote:
> 
> I'm having a networking problem with a UML installation.  I can boot the
> uml guest and get consoles and I can even get networking configured and
> working, but when I reboot the guest it complains that it can't find the
> ethernet device and /sbin/ifconfig shows that the ethernet device available
> is one greater than previously.  IOW, the first time, it's eth0, the second
> eth1, the third eth3, etc.  Obviously, this means
> that /etc/network/interfaces doesn't have the right eth device.  I can then
> manually configure the current device whether it is eth0 or eth22, but I'd
> rather not have to do that every time I boot the guest.  So my question is
> why does the device keep incrementing?  What controls it, and how can I
> make it stop?

It's probably the udev rules, try to find something like
/etc/udev/rules.d/z25_persistent-net.rules and persistent-net-generator
rules and remove them.

regards,
iustin

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 0 comments |
headers
Flavio | 4 Jul 2008 10:48
Picon
Gravatar

[uml-user] Problem running in SKAS4 Mode

Hello,

I've some problem running my UML in SKAS4 mode on my x86_64 (amd64) PC.
My host system has a 2.6.25-gentoo-r6 kernel with the SKAS4 patch for
2.6.25 kernels.
My UML kernel is 2.6.24.2 vanilla kernel with the skas4 patch for
2.6.24 kernels.

Why my guest is running in skas0 mode instead of skas4?

Should both the guest and the host kernel at the same version?

Regards,

Flavio

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 0 comments |
headers
David Fernández | 5 Jul 2008 19:27
Picon
Picon
Favicon

[uml-user] Running UML over Vmware

Hi,

I am making some tests running UML inside an Ubuntu-8.04-jeos running over Vmware ESX, but the performance
I get is very poor compared to running UML directly over a native Ubuntu. 

For example, starting a virtual machine takes almost three times more over ESX:

Ubuntu-jeos over VMware ESX in an Intel Xeon 3040 <at> 1.86GHz --> 32 secs

Ubuntu 8.04 Desktop over an Intel Core2 6400 <at> 2.13GHz --> 12 secs

The Ubuntu-jeos running over ESX has the VMware tools installed. 

Any hint about the cause of this big performance difference? Should the installation of the SKAS3/4 patch
on the host help to solve it? 

Thanks in advance and best regards;
David

P.D: guest machines are running kernel 2.6.24.3 and the host is running 2.6.24-19-virtual in the ESX host
and 2.6.24-17-generic in the other.

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 4 comments |
headers
Flavio | 6 Jul 2008 10:37
Picon
Gravatar

Re: [uml-user] Running UML over Vmware

2008/7/5 David Fernández <david <at> dit.upm.es>:
>
> Hi,
>
> I am making some tests running UML inside an Ubuntu-8.04-jeos running over Vmware ESX, but the
performance I get is very poor compared to running UML directly over a native Ubuntu.
>
> For example, starting a virtual machine takes almost three times more over ESX:
>
> Ubuntu-jeos over VMware ESX in an Intel Xeon 3040 <at> 1.86GHz --> 32 secs
>
> Ubuntu 8.04 Desktop over an Intel Core2 6400 <at> 2.13GHz --> 12 secs
>
> The Ubuntu-jeos running over ESX has the VMware tools installed.
>
> Any hint about the cause of this big performance difference? Should the installation of the SKAS3/4 patch
on the host help to solve it?
Hello,

try the skas4 patch for 2.6.24 kernel on both guest and host kernel
(skas3 is not longer supported).
Maybe you will have some performance increment. Note that most of the
times, virtualization my slow down guest processes. Using a
paravirtualization system (like XEN) may not decrease your guest
performances so much.

Regards,

Flavio

>
> Thanks in advance and best regards;
> David
>
> P.D: guest machines are running kernel 2.6.24.3 and the host is running 2.6.24-19-virtual in the ESX host
and 2.6.24-17-generic in the other.
>
>
>
> -------------------------------------------------------------------------
> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
> Studies have shown that voting for your favorite open source project,
> along with a healthy diet, reduces your potential for chronic lameness
> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
> _______________________________________________
> User-mode-linux-user mailing list
> User-mode-linux-user <at> lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/user-mode-linux-user

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 0 comments |
headers
Flavio | 6 Jul 2008 19:13
Picon
Gravatar

[uml-user] vanilla kernel 2.6.25.10: problem during compilation for the guest os

Hello everybody,

I'm experiencing problems during guest kernel compilation.
The vanilla kernel version is 2.6.25.10.

I've applied the skas4 patch for 2.6.25 kernels (without any error)
but I get the following error during compilation:

In file included from arch/um/os-Linux/start_up.c:29:
/opt/uml/src/linux-2.6.25.10/arch/um/include/skas/skas.h:29:22: error:
asm/user.h: No such file or directory
In file included from arch/um/os-Linux/start_up.c:29:
/opt/uml/src/linux-2.6.25.10/arch/um/include/skas/skas.h:75: error:
field 'fpregs' has incomplete type
make[1]: *** [arch/um/os-Linux/start_up.o] Error 1
make: *** [arch/um/os-Linux] Error 2

What's going on?

Regards,

Flavio

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 6 comments |
headers
Jeff Dike | 7 Jul 2008 18:37

Re: [uml-user] Running UML over Vmware

On Sat, Jul 05, 2008 at 07:27:44PM +0200, David Fernández wrote:
> Any hint about the cause of this big performance difference? Should
> the installation of the SKAS3/4 patch on the host help to solve it?  

skas4 will certainly help.  I get ~86% of native performance on a
kernel build with the latest skas4 patch.

       	     	      Jeff

--

-- 
Work email - jdike at linux dot intel dot com

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 2 comments |
headers
Nix | 9 Jul 2008 00:59
Picon

Re: [uml-user] Running UML over Vmware

On 7 Jul 2008, Jeff Dike told this:

> On Sat, Jul 05, 2008 at 07:27:44PM +0200, David Fernández wrote:
>> Any hint about the cause of this big performance difference? Should
>> the installation of the SKAS3/4 patch on the host help to solve it?  
>
> skas4 will certainly help.  I get ~86% of native performance on a
> kernel build with the latest skas4 patch.

I'll have to try it again one of these days. I miss skas3's speed,
and skas4 failed horribly last time I tried it...

-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
Permalink | Reply | 0 comments |

Gmane