9:12 a.m.
Hi,
In a oVirt 3.4.4 environnement, and using a
virt-p2v-0.9.1-2.20130730.1.el6.centos, I'm trying to convert a windows
2003 server 32 bits into an oVirt KVM image.
I already succeed doing that many times, but now, this is failing :
- All the usual process sounds good :
PXE -> DHCP -> connection to the P2V server -> choosing the profile
-> (checking the debug option) -> running the conversion
But in the end (100%), the only error message I get is :
"virt-v2v: Died at /usr/bin/virt-p2v-server line 411"
which does not help me that much, because it is related to some
conversion made by libguestfs, I suppose, and I correctly installed the
libguestfs-winsupport files into the correct directories.
An evidence of that is a recent failure when converting from KVM to
oVirt (so : _V_2V), SOLVED by installing the correct directories in the
correct places.
As I recently switched in this almost-new virt-p2p version (0.9.1), I
was able to activate some useful logs that are available there :
https://www.ecarnot.net/tmp/virt-p2v-server.1415622286.log.gz
It has been days since I'm trying to contact Matthew via IRC, but he
seems to love travelling away for it, and I pray he would find a minute
to read this and give a hint.
Thank you,
Best regards,
--
Nicolas Ecarnot
10:01 a.m.
On Mon, Nov 10, 2014 at 04:12:35PM +0100, Nicolas Ecarnot wrote:
Hi,
In a oVirt 3.4.4 environnement, and using a
virt-p2v-0.9.1-2.20130730.1.el6.centos, I'm trying to convert a
windows 2003 server 32 bits into an oVirt KVM image.
I already succeed doing that many times, but now, this is failing :
- All the usual process sounds good :
PXE -> DHCP -> connection to the P2V server -> choosing the profile
-> (checking the debug option) -> running the conversion
But in the end (100%), the only error message I get is :
"virt-v2v: Died at /usr/bin/virt-p2v-server line 411"
which does not help me that much, because it is related to some
conversion made by libguestfs, I suppose, and I correctly installed
the libguestfs-winsupport files into the correct directories.
An evidence of that is a recent failure when converting from KVM to
oVirt (so : _V_2V), SOLVED by installing the correct directories in
the correct places.
As I recently switched in this almost-new virt-p2p version (0.9.1),
I was able to activate some useful logs that are available there :
https://www.ecarnot.net/tmp/virt-p2v-server.1415622286.log.gz
It has been days since I'm trying to contact Matthew via IRC, but he
seems to love travelling away for it, and I pray he would find a
minute to read this and give a hint.
Thank you,
Best regards,
--
Nicolas Ecarnot
_______________________________________________
Libguestfs mailing list
Libguestfs(a)redhat.com
https://www.redhat.com/mailman/listinfo/libguestfs
Matt - any ideas?
There is also the new version of p2v/v2v available, although it
requires recent Fedora or RHEL/Centos 7.0.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine. Supports Linux and Windows.
http://people.redhat.com/~rjones/virt-df/
10:52 a.m.
On 10/11/14 16:01, Richard W.M. Jones wrote:
On Mon, Nov 10, 2014 at 04:12:35PM +0100, Nicolas Ecarnot wrote:
> Hi,
>
> In a oVirt 3.4.4 environnement, and using a
> virt-p2v-0.9.1-2.20130730.1.el6.centos, I'm trying to convert a
> windows 2003 server 32 bits into an oVirt KVM image.
>
> I already succeed doing that many times, but now, this is failing :
> - All the usual process sounds good :
> PXE -> DHCP -> connection to the P2V server -> choosing the profile
> -> (checking the debug option) -> running the conversion
>
> But in the end (100%), the only error message I get is :
> "virt-v2v: Died at /usr/bin/virt-p2v-server line 411"
> which does not help me that much, because it is related to some
> conversion made by libguestfs, I suppose, and I correctly installed
> the libguestfs-winsupport files into the correct directories.
>
> An evidence of that is a recent failure when converting from KVM to
> oVirt (so : _V_2V), SOLVED by installing the correct directories in
> the correct places.
>
> As I recently switched in this almost-new virt-p2p version (0.9.1),
> I was able to activate some useful logs that are available there :
> https://www.ecarnot.net/tmp/virt-p2v-server.1415622286.log.gz
>
> It has been days since I'm trying to contact Matthew via IRC, but he
> seems to love travelling away for it, and I pray he would find a
> minute to read this and give a hint.
>
> Thank you,
>
> Best regards,
>
> --
> Nicolas Ecarnot
>
> _______________________________________________
> Libguestfs mailing list
> Libguestfs(a)redhat.com
> https://www.redhat.com/mailman/listinfo/libguestfs
Matt - any ideas?
There is also the new version of p2v/v2v available, although it
requires recent Fedora or RHEL/Centos 7.0.
This is the last thing it does before unwinding:
libguestfs: trace: cp "/transferxdTiHj/windows/rhev-apt.exe" "/Program
Files/RedHat/V2V Firstboot"
libguestfs: trace: cp = 0
So, maybe the guest is out of disk space, or has filesystem corruption?
Matt
Rich.
--
Matthew Booth
Red Hat Engineering, Virtualisation Team
Phone: +442070094448 (UK)
GPG ID: D33C3490
GPG FPR: 3733 612D 2D05 5458 8A8A 1600 3441 EA19 D33C 3490
1:09 p.m.
On Mon, Nov 10, 2014 at 04:52:35PM +0000, Matthew Booth wrote:
On 10/11/14 16:01, Richard W.M. Jones wrote:
> On Mon, Nov 10, 2014 at 04:12:35PM +0100, Nicolas Ecarnot wrote:
>> Hi,
>>
>> In a oVirt 3.4.4 environnement, and using a
>> virt-p2v-0.9.1-2.20130730.1.el6.centos, I'm trying to convert a
>> windows 2003 server 32 bits into an oVirt KVM image.
>>
>> I already succeed doing that many times, but now, this is failing :
>> - All the usual process sounds good :
>> PXE -> DHCP -> connection to the P2V server -> choosing the profile
>> -> (checking the debug option) -> running the conversion
>>
>> But in the end (100%), the only error message I get is :
>> "virt-v2v: Died at /usr/bin/virt-p2v-server line 411"
>> which does not help me that much, because it is related to some
>> conversion made by libguestfs, I suppose, and I correctly installed
>> the libguestfs-winsupport files into the correct directories.
>>
>> An evidence of that is a recent failure when converting from KVM to
>> oVirt (so : _V_2V), SOLVED by installing the correct directories in
>> the correct places.
>>
>> As I recently switched in this almost-new virt-p2p version (0.9.1),
>> I was able to activate some useful logs that are available there :
>> https://www.ecarnot.net/tmp/virt-p2v-server.1415622286.log.gz
>>
>> It has been days since I'm trying to contact Matthew via IRC, but he
>> seems to love travelling away for it, and I pray he would find a
>> minute to read this and give a hint.
>>
>> Thank you,
>>
>> Best regards,
>>
>> --
>> Nicolas Ecarnot
>>
>> _______________________________________________
>> Libguestfs mailing list
>> Libguestfs(a)redhat.com
>> https://www.redhat.com/mailman/listinfo/libguestfs
>
> Matt - any ideas?
>
> There is also the new version of p2v/v2v available, although it
> requires recent Fedora or RHEL/Centos 7.0.
This is the last thing it does before unwinding:
libguestfs: trace: cp "/transferxdTiHj/windows/rhev-apt.exe" "/Program
Files/RedHat/V2V Firstboot"
libguestfs: trace: cp = 0
So, maybe the guest is out of disk space, or has filesystem corruption?
New v2v enforces a minimum free space on mountable fses which I don't
think was in the old version. If c:\ has less than 50 MB free it will
refuse to convert it. Use 'virt-df' to check.
That said I would be surprised if the 'cp' succeeded but v2v crashed
suddenly because of insufficient space.
Rich.
Matt
>
> Rich.
>
>
--
Matthew Booth
Red Hat Engineering, Virtualisation Team
Phone: +442070094448 (UK)
GPG ID: D33C3490
GPG FPR: 3733 612D 2D05 5458 8A8A 1600 3441 EA19 D33C 3490
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
libguestfs lets you edit virtual machines. Supports shell scripting,
bindings from many languages. http://libguestfs.org
1:28 p.m.
Le 10/11/2014 20:09, Richard W.M. Jones a écrit :
>> Matt - any ideas?
>>
>> There is also the new version of p2v/v2v available, although it
>> requires recent Fedora or RHEL/Centos 7.0.
>
> This is the last thing it does before unwinding:
>
> libguestfs: trace: cp "/transferxdTiHj/windows/rhev-apt.exe"
"/Program
> Files/RedHat/V2V Firstboot"
> libguestfs: trace: cp = 0
>
> So, maybe the guest is out of disk space, or has filesystem corruption?
New v2v enforces a minimum free space on mountable fses which I don't
think was in the old version. If c:\ has less than 50 MB free it will
refuse to convert it. Use 'virt-df' to check.
That said I would be surprised if the 'cp' succeeded but v2v crashed
suddenly because of insufficient space.
Hello,
Do I have to understand that :
- I should perform a fsck on my guest filesystem before the P2V?
- I should ensure there is enough available space on it?
The partition I'm trying to convert is 50Go large, filled up with
something like 40Go. Even with a big error margin, it seems comfortable.
Before the P2V, I did no fsck, but I used to microsoft defrag tool (the
old 2003 one). I'm not sure it also checks the filesystem integrity...
- Finally, when you write about the "NEW" v2v, what version do you speak
about? Is the one I used (0.9.1.2) not new enough? (I know it is not the
very last, but quite decent though).
Well, thank you for having answered.
On this conversion, I have no choice but succeed.
I also tried the old livecd+netcat+dd option, but the OS did not even
boot, even if I copied the MBR and double check the special flag with an
hex editor ("FF").
So I will keep trying until success.
--
Nicolas Ecarnot
5:10 p.m.
On Mon, Nov 10, 2014 at 08:28:53PM +0100, Nicolas Ecarnot wrote:
Le 10/11/2014 20:09, Richard W.M. Jones a écrit :
>>>Matt - any ideas?
>>>
>>>There is also the new version of p2v/v2v available, although it
>>>requires recent Fedora or RHEL/Centos 7.0.
>>
>>This is the last thing it does before unwinding:
>>
>>libguestfs: trace: cp "/transferxdTiHj/windows/rhev-apt.exe"
"/Program
>>Files/RedHat/V2V Firstboot"
>>libguestfs: trace: cp = 0
>>
>>So, maybe the guest is out of disk space, or has filesystem corruption?
>
>New v2v enforces a minimum free space on mountable fses which I don't
>think was in the old version. If c:\ has less than 50 MB free it will
>refuse to convert it. Use 'virt-df' to check.
>
>That said I would be surprised if the 'cp' succeeded but v2v crashed
>suddenly because of insufficient space.
Hello,
Do I have to understand that :
- I should perform a fsck on my guest filesystem before the P2V?
- I should ensure there is enough available space on it?
The partition I'm trying to convert is 50Go large, filled up with
something like 40Go. Even with a big error margin, it seems
comfortable.
Before the P2V, I did no fsck, but I used to microsoft defrag tool
(the old 2003 one). I'm not sure it also checks the filesystem
integrity...
- Finally, when you write about the "NEW" v2v, what version do you
speak about? Is the one I used (0.9.1.2) not new enough? (I know it
is not the very last, but quite decent though).
The new version is 1.28, part of libguestfs. See:
https://www.redhat.com/archives/libguestfs/2014-May/msg00090.html
Rich.
Well, thank you for having answered.
On this conversion, I have no choice but succeed.
I also tried the old livecd+netcat+dd option, but the OS did not
even boot, even if I copied the MBR and double check the special
flag with an hex editor ("FF").
So I will keep trying until success.
--
Nicolas Ecarnot
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine. Supports Linux and Windows.
http://people.redhat.com/~rjones/virt-df/
9:06 a.m.
On Mon, Nov 10, 2014 at 08:28:53PM +0100, Nicolas Ecarnot wrote:
Le 10/11/2014 20:09, Richard W.M. Jones a écrit :
>>>Matt - any ideas?
>>>
>>>There is also the new version of p2v/v2v available, although it
>>>requires recent Fedora or RHEL/Centos 7.0.
>>
>>This is the last thing it does before unwinding:
>>
>>libguestfs: trace: cp "/transferxdTiHj/windows/rhev-apt.exe"
"/Program
>>Files/RedHat/V2V Firstboot"
>>libguestfs: trace: cp = 0
>>
>>So, maybe the guest is out of disk space, or has filesystem corruption?
>
>New v2v enforces a minimum free space on mountable fses which I don't
>think was in the old version. If c:\ has less than 50 MB free it will
>refuse to convert it. Use 'virt-df' to check.
>
>That said I would be surprised if the 'cp' succeeded but v2v crashed
>suddenly because of insufficient space.
Hello,
Do I have to understand that :
- I should perform a fsck on my guest filesystem before the P2V?
Not AFAIK.
- I should ensure there is enough available space on it?
There needs to be some space, but probably not very much. If the root
filesystem has no free space at all, then virt-v2v cannot install the
packages and files that it needs.
The partition I'm trying to convert is 50Go large, filled up
with
something like 40Go. Even with a big error margin, it seems
comfortable.
This is plenty.
Before the P2V, I did no fsck, but I used to microsoft defrag tool
(the old 2003 one). I'm not sure it also checks the filesystem
integrity...
- Finally, when you write about the "NEW" v2v, what version do you
speak about? Is the one I used (0.9.1.2) not new enough? (I know it
is not the very last, but quite decent though).
1.28.x here:
https://www.redhat.com/archives/libguestfs/2014-May/msg00090.html
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
libguestfs lets you edit virtual machines. Supports shell scripting,
bindings from many languages. http://libguestfs.org
12:45 p.m.
Hello,
Answers below :
Le 13/11/2014 16:06, Richard W.M. Jones a écrit :
>>> This is the last thing it does before unwinding:
>>>
>>> libguestfs: trace: cp "/transferxdTiHj/windows/rhev-apt.exe"
"/Program
>>> Files/RedHat/V2V Firstboot"
>>> libguestfs: trace: cp = 0
>>>
>>> So, maybe the guest is out of disk space, or has filesystem corruption?
>>
>> New v2v enforces a minimum free space on mountable fses which I don't
>> think was in the old version. If c:\ has less than 50 MB free it will
>> refuse to convert it. Use 'virt-df' to check.
>>
>> That said I would be surprised if the 'cp' succeeded but v2v crashed
>> suddenly because of insufficient space.
>
> Hello,
>
> Do I have to understand that :
> - I should perform a fsck on my guest filesystem before the P2V?
Not AFAIK.
Well, anyway, on the source, I forced a fsck, then tried again, with no
improvement.
> - Finally, when you write about the "NEW" v2v, what
version do you
> speak about? Is the one I used (0.9.1.2) not new enough? (I know it
> is not the very last, but quite decent though).
1.28.x here:
https://www.redhat.com/archives/libguestfs/2014-May/msg00090.html
I am using libguestfs through oVirt 3.4 (RHEV), and whoever knows this
project knows how intricated the packages are.
As this is a production environment, I can not play with trying to
change libguestfs on the oVirt manager.
But I would be enthusiastic to install a dedicated P2V server, on RH7,
with the very last libguestfs, if that helps me to insure our future P2V
will succeed.
But according to what I understood so far, I'm not sure one can install
such a P2V server, independent of any oVirt setup.
As a matter of fact, I'll be glad to test the last libguestfs version,
but I don't know where I could use it?
--
Nicolas Ecarnot
1:31 p.m.
On Mon, Nov 17, 2014 at 07:45:12PM +0100, Nicolas Ecarnot wrote:
But I would be enthusiastic to install a dedicated P2V server, on
RH7, with the very last libguestfs, if that helps me to insure our
future P2V will succeed.
But according to what I understood so far, I'm not sure one can
install such a P2V server, independent of any oVirt setup.
As a matter of fact, I'll be glad to test the last libguestfs
version, but I don't know where I could use it?
Just install a RHEL 7.0 VM, and upgrade it with the packages by
following the instructions here:
https://www.redhat.com/archives/libguestfs/2014-May/msg00090.html
The other minimum requirements are covered here:
http://libguestfs.org/virt-v2v.1.html#resource-requirements
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
libguestfs lets you edit virtual machines. Supports shell scripting,
bindings from many languages. http://libguestfs.org
8:06 a.m.
Le 17/11/2014 20:31, Richard W.M. Jones a écrit :
On Mon, Nov 17, 2014 at 07:45:12PM +0100, Nicolas Ecarnot wrote:
> But I would be enthusiastic to install a dedicated P2V server, on
> RH7, with the very last libguestfs, if that helps me to insure our
> future P2V will succeed.
> But according to what I understood so far, I'm not sure one can
> install such a P2V server, independent of any oVirt setup.
>
> As a matter of fact, I'll be glad to test the last libguestfs
> version, but I don't know where I could use it?
Just install a RHEL 7.0 VM, and upgrade it with the packages by
following the instructions here:
https://www.redhat.com/archives/libguestfs/2014-May/msg00090.html
The other minimum requirements are covered here:
http://libguestfs.org/virt-v2v.1.html#resource-requirements
Rich.
Richard, Matthew et al,
In a VM, I installed a RHEL 7.0 and tried to install libguestfs-tools
but I saw there were many errors related to missing repositories and
other errors.
I wasn't able to cope with these issues.
So I tried to install another VM with CentOS 7.0, and I tried to install
libguestfs-tools.
Another problem came from the too old version of glusterfs-api, so I
added the lastest glusterfs repo definition, upgraded, and succeed in
installing libguestfs-tools and virt-v2v.
Then I PXE-boot my source physical machine (well, at present, at test
time, another VM but whatever), when getting the graphical prompt to
connect to the conversion server, I get
"Failed to start virt-v2v-server on remote server".
though I witness the SSH connection is OK on the conversion server.
On this conversion server, I find no trace of any
/usr/bin/virt-p2v-server binary.
See :
[root@serv-p2v-adm1 log]# rpm -qa virt-v2v
virt-v2v-1.28.1-1.11.el7.x86_64
[root@serv-p2v-adm1 log]# rpm -qa virt-v2v
virt-v2v-1.28.1-1.11.el7.x86_64
[root@serv-p2v-adm1 log]# rpm -ql virt-v2v
/usr/bin/virt-p2v-make-disk
/usr/bin/virt-p2v-make-kickstart
/usr/bin/virt-v2v
/usr/libexec/virt-p2v
/usr/share/doc/virt-v2v-1.28.1
/usr/share/doc/virt-v2v-1.28.1/COPYING
/usr/share/doc/virt-v2v-1.28.1/README
/usr/share/doc/virt-v2v-1.28.1/TODO
/usr/share/man/man1/virt-p2v-make-disk.1.gz
/usr/share/man/man1/virt-p2v-make-kickstart.1.gz
/usr/share/man/man1/virt-p2v.1.gz
/usr/share/man/man1/virt-v2v.1.gz
/usr/share/virt-p2v
/usr/share/virt-p2v/issue
/usr/share/virt-p2v/launch-virt-p2v
/usr/share/virt-p2v/p2v.ks.in
/usr/share/virt-p2v/p2v.service
/usr/share/virt-tools
/usr/share/virt-tools/rhev-apt.exe
/usr/share/virt-tools/rhsrvany.exe
Is this virt-p2v-server really missing, or is it done another way now
we're on 7.x ?
I ask that because when comparing with the same rpm on a 6.6, it
contains much more things, amongst them the binary server and the
numerous perl scripts.
--
Nicolas Ecarnot
9:11 a.m.
On Tue, Nov 18, 2014 at 03:06:44PM +0100, Nicolas Ecarnot wrote:
[...]
Then I PXE-boot my source physical machine (well, at present, at
test time, another VM but whatever), when getting the graphical
prompt to connect to the conversion server, I get
"Failed to start virt-v2v-server on remote server".
though I witness the SSH connection is OK on the conversion server.
On this conversion server, I find no trace of any
/usr/bin/virt-p2v-server binary.
You're using the wrong version of virt-p2v. You have to use the new
version.
You can build it yourself using the scripts 'virt-p2v-make-disk' or
'virt-p2v-make-kickstart' (read the manual pages first), or try a
recent version that I put here:
http://oirase.annexia.org/virt-p2v/
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from scratch
http://libguestfs.org/virt-builder.1.html
2:08 a.m.
Le 18/11/2014 16:11, Richard W.M. Jones a écrit :
On Tue, Nov 18, 2014 at 03:06:44PM +0100, Nicolas Ecarnot wrote:
[...]
> Then I PXE-boot my source physical machine (well, at present, at
> test time, another VM but whatever), when getting the graphical
> prompt to connect to the conversion server, I get
> "Failed to start virt-v2v-server on remote server".
> though I witness the SSH connection is OK on the conversion server.
>
> On this conversion server, I find no trace of any
> /usr/bin/virt-p2v-server binary.
You're using the wrong version of virt-p2v. You have to use the new
version.
You can build it yourself using the scripts 'virt-p2v-make-disk' or
'virt-p2v-make-kickstart' (read the manual pages first), or try a
recent version that I put here:
http://oirase.annexia.org/virt-p2v/
Rich.
Richard,
Thanks for having posted your iso (and the .tar).
I installed the .tar to replace my /tftpboot/p2vboot directory, but when
booting, I'm hit by the now know bug "Failed to load ldlinux.c32"
(PXELINUX 6.03)
According to what I googled (quite intensively), it may be due to the
fact that the files contained in /tftpboot/p2vboot are not at the root
(/tftpboot).
Before going further, I have to say that I still need to succeed in this
P2V task.
Since long, I'm still very motivated to try and keep progressing.
But today, failures after failures, I must admit I'm getting a little
tired of all these steps :
- tftpboot install
- pxe setup
- dhcp setup with special options
- p2vboot of the source machine
- failure in the perl script with no explanation
- install of the updated version of virt-p2v.iso, to get debug checkbox
- fsck of the source
- no real help from the debug log (a "cp" command failing, but why?)
- install of a more recent virt-p2v, that can not boot due to
ldlinux.c32 missing
As the ldlinux.c32 file was missing, I found it somewhere and placed it
into /tftpboot, then into /tftpboot/p2vboot, but it is not found either
after gPXE find it ip address.
I tried to use the virt-p2v-make-disk, but it failed to connect to a
libvirt socket (obviously, there is none running in this VM.)
Well, I take one day off (this issue). I'll go breathe another air, and
see you tomorrow because I'm getting a point I find it necessary.
Thanks for reading this far. See you tomorow.
--
Nicolas Ecarnot
3:47 a.m.
On Wed, Nov 19, 2014 at 09:08:36AM +0100, Nicolas Ecarnot wrote:
I tried to use the virt-p2v-make-disk, but it failed to connect to a
libvirt socket (obviously, there is none running in this VM.)
To avoid libvirt problems, do:
export LIBGUESTFS_BACKEND=direct
Why don't you use the ISO? Burn it onto a CD and boot it.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-top is 'top' for virtual machines. Tiny program with many
powerful monitoring features, net stats, disk stats, logging, etc.
http://people.redhat.com/~rjones/virt-top
5:58 a.m.
Le 19/11/2014 10:47, Richard W.M. Jones a écrit :
On Wed, Nov 19, 2014 at 09:08:36AM +0100, Nicolas Ecarnot wrote:
> I tried to use the virt-p2v-make-disk, but it failed to connect to a
> libvirt socket (obviously, there is none running in this VM.)
To avoid libvirt problems, do:
export LIBGUESTFS_BACKEND=direct
Why don't you use the ISO? Burn it onto a CD and boot it.
Rich.
OK Richard,
I promised I would wait a day before getting back to this issue, but you
know how to tease a sysadmin ;)
Well, I did boot your iso file as a CD, and now it is booting and I get
the very last version of virt-p2v (Wow, lots of additionnal options :
great!).
I can confirm the ssh test connexion is useful and always working.
I like the new fields I can fill to (try to) direct the flow towards a
real oVirt/RHEV NFS storage domain.
But it is now failing with an error message telling it can not connect
to the local socket /var/run/libvirt/libvirt-sock.
I guess this is related to your previous answer :
On 19/11/2014 10:47, Richard W.M. Jones wrote :
On Wed, Nov 19, 2014 at 09:08:36AM +0100, Nicolas Ecarnot wrote:
> I tried to use the virt-p2v-make-disk, but it failed to connect to a
> libvirt socket (obviously, there is none running in this VM.)
To avoid libvirt problems, do:
export LIBGUESTFS_BACKEND=direct
But in this use case, I don't know where to set this variable?
I'd be glad to set it on the source machine, before running the virt-p2v
GUI. But the usual alt-F1 or alt-F2 does not give me a shell prompt
anymore. Instead, a quickly blinking message shows it is trying to
(systemctl) start the service and fails.
I'd be glad to set it in the conversion server, but where and when?
In case this could be useful, here is below the versions implied :
- virt-v2v: libguestfs 1.28.1 (x86_64)
- virt-p2v-20141119
At this point, I guess I have something to do on the conversion server...?
--
Nicolas Ecarnot
6:03 a.m.
On Wed, Nov 19, 2014 at 12:58:11PM +0100, Nicolas Ecarnot wrote:
But it is now failing with an error message telling it can not
connect to the local socket /var/run/libvirt/libvirt-sock.
This is basically a libvirt bug that you have to solve separately.
I'm guessing that basic commands such as `virsh nodeinfo' also fail.
I don't know what the reason for that will be, but you can ask the
libvirt users list to help you fix it.
> To avoid libvirt problems, do:
>
> export LIBGUESTFS_BACKEND=direct
But in this use case, I don't know where to set this variable?
I believe it will work if you put this into root's .bash_profile or
.bashrc (ie. /root/.bash_profile).
Given the amount of trouble that libvirt causes us, it's a reasonable
feature request to add a tick box or something to allow the backend to
be adjustable from the GUI. See the FAQ for how to file feature
requests.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from scratch
http://libguestfs.org/virt-builder.1.html
7:41 a.m.
Le 19/11/2014 13:03, Richard W.M. Jones a écrit :
On Wed, Nov 19, 2014 at 12:58:11PM +0100, Nicolas Ecarnot wrote:
> But it is now failing with an error message telling it can not
> connect to the local socket /var/run/libvirt/libvirt-sock.
This is basically a libvirt bug that you have to solve separately.
I'm guessing that basic commands such as `virsh nodeinfo' also fail.
I don't know what the reason for that will be, but you can ask the
libvirt users list to help you fix it.
>> To avoid libvirt problems, do:
>>
>> export LIBGUESTFS_BACKEND=direct
>
> But in this use case, I don't know where to set this variable?
I believe it will work if you put this into root's .bash_profile or
.bashrc (ie. /root/.bash_profile).
PROGRESS !
I added the setting above into the root profile, and now, things are
beginning to move.
The first server I have to convert is a production server, so I can not
easily halt it to run many test conversion.
So these days I'm trying to convert a VM (and I don't know if this is
possible though I don't see why it wouldn't be).
When trying to p2v an Oracle Linux 7.0 VM, it ran very verbosely then
failed saying it was not a kind of OS it knows how to convert.
When trying to p2v a old windows XP VM, it ran the same long way then
failed telling that no root device was found in this operating system image.
As I read in detail the very long log, I saw different things failing,
and amongst them, the issue telling that the 36:36 permissions on the
NFS export share will prevent the import into RHEV. I don't know if it
is an issue at this point. Anyway, I don't see any image coming into
this nfs subtree at present.
Asap, I will try to run a P2V on the windows 2003 server source - the
one I'm struggling for for weeks, and let you know.
Stay tuned.
PS : Thank you Richard for your support.
--
Nicolas Ecarnot
8:03 a.m.
On Wed, Nov 19, 2014 at 02:41:49PM +0100, Nicolas Ecarnot wrote:
When trying to p2v an Oracle Linux 7.0 VM, it ran very verbosely
then failed saying it was not a kind of OS it knows how to convert.
We could probably support OEL with a bit of effort, but it's not
supported now. The virt-v2v manual (which you should read, as well as
the virt-p2v manual) provides a list of all OSes which are supported.
When trying to p2v a old windows XP VM, it ran the same long way
then failed telling that no root device was found in this operating
system image.
Windows XP should work. If it doesn't work you have to enable
debugging (this is now enabled by default in all new virt-p2v
versions), and then provide the full debug logs. See the instructions
for where these are located on the conversion server.
As I read in detail the very long log, I saw different things
failing, and amongst them, the issue telling that the 36:36
permissions on the NFS export share will prevent the import into
RHEV. I don't know if it is an issue at this point. Anyway, I don't
see any image coming into this nfs subtree at present.
When using '-o rhev', permissions should be set correctly for you
automatically. Instead of summarising and interpreting what the debug
logs say (which is not useful), provide the debug logs.
However if it doesn't succeed in the conversion, then it's not going
to appear in RHEV.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-p2v converts physical machines to virtual machines. Boot with a
live CD or over the network (PXE) and turn machines into KVM guests.
http://libguestfs.org/virt-v2v
8:50 a.m.
Le 19/11/2014 15:03, Richard W.M. Jones a écrit :
On Wed, Nov 19, 2014 at 02:41:49PM +0100, Nicolas Ecarnot wrote:
> When trying to p2v an Oracle Linux 7.0 VM, it ran very verbosely
> then failed saying it was not a kind of OS it knows how to convert.
We could probably support OEL with a bit of effort, but it's not
supported now. The virt-v2v manual (which you should read, as well as
the virt-p2v manual) provides a list of all OSes which are supported.
Richard,
Don't worry, I can ensure you that I'm reading and reading and reading
again.
BTW, these docs are very well done. This is comfortable.
Concerning OL7, this is not a serious issue, this was just a test.
> When trying to p2v a old windows XP VM, it ran the same long way
> then failed telling that no root device was found in this operating
> system image.
Windows XP should work. If it doesn't work you have to enable
debugging (this is now enabled by default in all new virt-p2v
versions), and then provide the full debug logs. See the instructions
for where these are located on the conversion server.
In the mean time, I googled around and found that though I copied the
right virtio-win hierarchy into the conversion server, I also had to
install the missing libguestfs-winsupport.
Another run of p2v of the previous windows XP succeed!
I did not try to import it into RHEV, but I can see it appearing in the
export domain list. So I guess the hardest part is done.
> As I read in detail the very long log, I saw different things
> failing, and amongst them, the issue telling that the 36:36
> permissions on the NFS export share will prevent the import into
> RHEV. I don't know if it is an issue at this point. Anyway, I don't
> see any image coming into this nfs subtree at present.
When using '-o rhev', permissions should be set correctly for you
automatically. Instead of summarising and interpreting what the debug
logs say (which is not useful), provide the debug logs.
I knew you would say that, but I'm still searching for a way to
copy-paste from a Dell iDrac environnment...
Well, now, I had the opportunity to run the real P2V on the 2003 server,
and it is failing.
But it appears that the partition seems more complex than what I'm used
to see ("two system (dell?) or windows system partitions").
I choose the third one, that is the one containing the system (C:) and
marked as bootable.
It is failing very quickly, with the following error :
--------------------
Creating an overlay to protect the source blahblah
qemu-img create -q -f qcow2 -b 'nbd:localhost:55756' -o
'compat=1.1,backing_fmt=raw' /var/tmp/v2vblahblah.qcow2
nbd.c:nbd_receive_negociate():L501: read failed
qemu-img: /var/tmp/v2vvovblahbla.qcow2: could not open
'nbd:localhost:55756': could not open image: invalid argument: invalid
argument
then other blahblahblah before closing.
I'll try another run by selecting every partition.
--
Nicolas Ecarnot
9:33 a.m.
On Wed, Nov 19, 2014 at 03:50:24PM +0100, Nicolas Ecarnot wrote:
Well, now, I had the opportunity to run the real P2V on the 2003
server, and it is failing.
But it appears that the partition seems more complex than what I'm
used to see ("two system (dell?) or windows system partitions").
I choose the third one, that is the one containing the system (C:)
and marked as bootable.
It is failing very quickly, with the following error :
--------------------
Creating an overlay to protect the source blahblah
qemu-img create -q -f qcow2 -b 'nbd:localhost:55756' -o
'compat=1.1,backing_fmt=raw' /var/tmp/v2vblahblah.qcow2
nbd.c:nbd_receive_negociate():L501: read failed
qemu-img: /var/tmp/v2vvovblahbla.qcow2: could not open
'nbd:localhost:55756': could not open image: invalid argument:
invalid argument
then other blahblahblah before closing.
You need to post the complete debug log. It's in /tmp on the
conversion server.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
libguestfs lets you edit virtual machines. Supports shell scripting,
bindings from many languages. http://libguestfs.org
9:42 a.m.
Le 19/11/2014 16:33, Richard W.M. Jones a écrit :
You need to post the complete debug log. It's in /tmp on the
conversion server.
Ah nice, I ignored that. Here it is:
virt-v2v: libguestfs 1.28.1 (x86_64)
[ 0.0] Opening the source -i libvirtxml
/tmp/virt-p2v-20141119-z1eg0fua/physical.xml
libvirt xml is:
<?xml version="1.0"?>
<!-- NOTE!
This libvirt XML is generated by the virt-p2v front end, in
order to communicate with the backend virt-v2v process running
on the conversion server. It is a minimal description of the
physical machine. If the target of the conversion is libvirt,
then virt-v2v will generate the real target libvirt XML, which
has only a little to do with the XML in this file.
TL;DR: Don't try to load this XML into libvirt. -->
<domain type="physical">
<name>bonjour</name>
<memory unit="KiB">4194304</memory>
<currentMemory unit="KiB">4194304</currentMemory>
<vcpu>2</vcpu>
<os>
<type arch="x86_64">hvm</type>
</os>
<features>
<acpi/>
<apic/>
<pae/>
</features>
<devices>
<disk type="network" device="disk">
<driver name="qemu" type="raw"/>
<source protocol="nbd">
<host name="localhost" port="33363"/>
</source>
<target dev="sda"/>
</disk>
<disk type="network" device="disk">
<driver name="qemu" type="raw"/>
<source protocol="nbd">
<host name="localhost" port="49584"/>
</source>
<target dev="sdb"/>
</disk>
<disk type="network" device="disk">
<driver name="qemu" type="raw"/>
<source protocol="nbd">
<host name="localhost" port="41988"/>
</source>
<target dev="sdc"/>
</disk>
</devices>
</domain>
source name: bonjour
hypervisor type: physical
memory: 4294967296 (bytes)
nr vCPUs: 2
CPU features: pae,apic,acpi
display:
disks:
nbd:localhost:33363 (raw) [sda]
nbd:localhost:49584 (raw) [sdb]
nbd:localhost:41988 (raw) [sdc]
removable media:
NICs:
[ 0.0] Creating an overlay to protect the source from being modified
qemu-img create -q -f qcow2 -b 'nbd:localhost:33363' -o
'compat=1.1,backing_fmt=raw' /var/tmp/v2vovl636444.qcow2
nbd.c:nbd_receive_negotiate():L501: read failed
qemu-img: /var/tmp/v2vovl636444.qcow2: Could not open
'nbd:localhost:33363': Could not open image: Invalid argument: Invalid
argument
virt-v2v: error: qemu-img command failed, see earlier errors
If reporting bugs, run virt-v2v with debugging enabled and include the
complete output:
virt-v2v -v -x [...]
--
Nicolas Ecarnot
11:12 a.m.
On Wed, Nov 19, 2014 at 04:42:23PM +0100, Nicolas Ecarnot wrote:
disks:
nbd:localhost:33363 (raw) [sda]
nbd:localhost:49584 (raw) [sdb]
nbd:localhost:41988 (raw) [sdc]
removable media:
NICs:
[ 0.0] Creating an overlay to protect the source from being modified
qemu-img create -q -f qcow2 -b 'nbd:localhost:33363' -o
'compat=1.1,backing_fmt=raw' /var/tmp/v2vovl636444.qcow2
nbd.c:nbd_receive_negotiate():L501: read failed
qemu-img: /var/tmp/v2vovl636444.qcow2: Could not open
'nbd:localhost:33363': Could not open image: Invalid argument:
Invalid argument
What is supposed to happen is that qemu-nbd is started (3 times) on
the virt-p2v machine. Then 3 x ssh connections are made which reverse
proxy the nbd connection from virt-v2v back to virt-p2v, the
equivalent of:
ssh -R 33363:localhost:<hostport> virt-v2v-server
Somehow that's not happening.
I'm not sure what. Is this reproducible? Do you have an unusual
firewall configuration on the virt-v2v conversion server?
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW
3:05 p.m.
Le 19/11/2014 18:12, Richard W.M. Jones a écrit :
On Wed, Nov 19, 2014 at 04:42:23PM +0100, Nicolas Ecarnot wrote:
> disks:
> nbd:localhost:33363 (raw) [sda]
> nbd:localhost:49584 (raw) [sdb]
> nbd:localhost:41988 (raw) [sdc]
> removable media:
>
> NICs:
>
> [ 0.0] Creating an overlay to protect the source from being modified
> qemu-img create -q -f qcow2 -b 'nbd:localhost:33363' -o
> 'compat=1.1,backing_fmt=raw' /var/tmp/v2vovl636444.qcow2
> nbd.c:nbd_receive_negotiate():L501: read failed
> qemu-img: /var/tmp/v2vovl636444.qcow2: Could not open
> 'nbd:localhost:33363': Could not open image: Invalid argument:
> Invalid argument
What is supposed to happen is that qemu-nbd is started (3 times) on
the virt-p2v machine. Then 3 x ssh connections are made which reverse
proxy the nbd connection from virt-v2v back to virt-p2v, the
equivalent of:
ssh -R 33363:localhost:<hostport> virt-v2v-server
Somehow that's not happening.
I'm not sure what. Is this reproducible? Do you have an unusual
firewall configuration on the virt-v2v conversion server?
Rich.
It is 100% reproductible.
Your interesting comment lead me to keep a better look at
/var/log/secure on the conversion server :
The "test" button succeed in connecting. I see it in the logs.
Nov 19 21:48:23 serv-p2v-adm1 sshd[2654]: Accepted password for root
from 192.168.49.5 port 36574 ssh2
Nov 19 21:48:23 serv-p2v-adm1 sshd[2654]: pam_unix(sshd:session):
session opened for user root by (uid=0)
Nov 19 21:48:23 serv-p2v-adm1 sshd[2654]: Received disconnect from
192.168.49.5: 11: disconnected by user
Nov 19 21:48:23 serv-p2v-adm1 sshd[2654]: pam_unix(sshd:session):
session closed for user root
Here I wait on purpose a long time to distinguish what is happening in
the log files.
I'm filling in the different fields in the conversion screen, the I
launch the conversion, and I see that :
(192.168.49.5 is the source machine)
Nov 19 21:54:52 serv-p2v-adm1 sshd[2725]: Accepted password for root
from 192.168.49.5 port 36575 ssh2
Nov 19 21:54:52 serv-p2v-adm1 sshd[2725]: pam_unix(sshd:session):
session opened for user root by (uid=0)
Nov 19 21:54:52 serv-p2v-adm1 sshd[2727]: Accepted password for root
from 192.168.49.5 port 36576 ssh2
Nov 19 21:54:52 serv-p2v-adm1 sshd[2727]: pam_unix(sshd:session):
session opened for user root by (uid=0)
Nov 19 21:54:53 serv-p2v-adm1 sshd[2725]: channel 2: open failed:
connect failed: Connection refused
Nov 19 21:54:53 serv-p2v-adm1 sshd[2727]: Received disconnect from
192.168.49.5: 11: disconnected by user
Nov 19 21:54:53 serv-p2v-adm1 sshd[2725]: Received disconnect from
192.168.49.5: 11: disconnected by user
Nov 19 21:54:53 serv-p2v-adm1 sshd[2725]: pam_unix(sshd:session):
session closed for user root
Nov 19 21:54:53 serv-p2v-adm1 sshd[2727]: pam_unix(sshd:session):
session closed for user root
Indeed, I see "Connection refused".
Let's investigate on this.
On the conversion server (serv-p2v-adm1), before running all this, I did
a "iptables -F" (and a restart of NFS - not knowing if it was needed),
so I have no rule in iptables now.
Do I have to find a way to increase the verbosity of sshd in
/var/log/secure?
Another point : as I wrote it before, I have no way to get a shell
prompt on the source machine, because alt-F1 leads to a blinking screen
trying to run virt-p2v and failing and respawning.
I nmap'd this source host, and saw that ssh was open.
In order to run the action step by step, is there a way to connect to
this root account? What password? (already tried
secret/p2v/root/password/PASSWORD/redhatrulez but that did not the trick)
Regards,
--
Nicolas Ecarnot
3:19 p.m.
On Wed, Nov 19, 2014 at 10:05:53PM +0100, Nicolas Ecarnot wrote:
Nov 19 21:54:52 serv-p2v-adm1 sshd[2727]: pam_unix(sshd:session):
session opened for user root by (uid=0)
Nov 19 21:54:53 serv-p2v-adm1 sshd[2725]: channel 2: open failed:
connect failed: Connection refused
Nov 19 21:54:53 serv-p2v-adm1 sshd[2727]: Received disconnect from
192.168.49.5: 11: disconnected by user
Nov 19 21:54:53 serv-p2v-adm1 sshd[2725]: Received disconnect from
192.168.49.5: 11: disconnected by user
[...]
Indeed, I see "Connection refused".
Let's investigate on this.
On the conversion server (serv-p2v-adm1), before running all this, I
did a "iptables -F" (and a restart of NFS - not knowing if it was
needed), so I have no rule in iptables now.
Do I have to find a way to increase the verbosity of sshd in
/var/log/secure?
You might want to check that sshd allows port forwarding
('AllowTcpForwarding yes' in /etc/ssh/sshd_config).
It would be useful to know what the actual sshd failure is.
This is how ssh is invoked on the virt-p2v side:
https://github.com/libguestfs/libguestfs/blob/master/p2v/ssh.c#L174
plus
https://github.com/libguestfs/libguestfs/blob/master/p2v/ssh.c#L595
Another point : as I wrote it before, I have no way to get a shell
prompt on the source machine, because alt-F1 leads to a blinking
screen trying to run virt-p2v and failing and respawning.
I nmap'd this source host, and saw that ssh was open.
In order to run the action step by step, is there a way to connect
to this root account? What password? (already tried
secret/p2v/root/password/PASSWORD/redhatrulez but that did not the
trick)
I'm fairly sure the root account is disabled, and sshd isn't supposed
to be running, but apparently it is (a bug). It's also a bug that
there is no shell on another other virtual console, but TBH I don't
know how to fix that ...
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-top is 'top' for virtual machines. Tiny program with many
powerful monitoring features, net stats, disk stats, logging, etc.
http://people.redhat.com/~rjones/virt-top
3:44 p.m.
Le 19/11/2014 22:19, Richard W.M. Jones a écrit :
On Wed, Nov 19, 2014 at 10:05:53PM +0100, Nicolas Ecarnot wrote:
> Nov 19 21:54:52 serv-p2v-adm1 sshd[2727]: pam_unix(sshd:session):
> session opened for user root by (uid=0)
> Nov 19 21:54:53 serv-p2v-adm1 sshd[2725]: channel 2: open failed:
> connect failed: Connection refused
> Nov 19 21:54:53 serv-p2v-adm1 sshd[2727]: Received disconnect from
> 192.168.49.5: 11: disconnected by user
> Nov 19 21:54:53 serv-p2v-adm1 sshd[2725]: Received disconnect from
> 192.168.49.5: 11: disconnected by user
[...]
>
> Indeed, I see "Connection refused".
> Let's investigate on this.
>
> On the conversion server (serv-p2v-adm1), before running all this, I
> did a "iptables -F" (and a restart of NFS - not knowing if it was
> needed), so I have no rule in iptables now.
>
> Do I have to find a way to increase the verbosity of sshd in
> /var/log/secure?
You might want to check that sshd allows port forwarding
('AllowTcpForwarding yes' in /etc/ssh/sshd_config).
This is the next thing I tried after posting my previous reply.
It was commented, and the man page says it is enable by default.
Anyway, I forced it on yes, restarted sshd and tried again but with no
success.
I'm trying the same operation after increasing the sshd log verbosity.
--
Nicolas Ecarnot
4:47 p.m.
Le 19/11/2014 22:44, Nicolas Ecarnot a écrit :
>> Indeed, I see "Connection refused".
>> Let's investigate on this.
>>
>> On the conversion server (serv-p2v-adm1), before running all this, I
>> did a "iptables -F" (and a restart of NFS - not knowing if it was
>> needed), so I have no rule in iptables now.
>>
>> Do I have to find a way to increase the verbosity of sshd in
>> /var/log/secure?
>
> You might want to check that sshd allows port forwarding
> ('AllowTcpForwarding yes' in /etc/ssh/sshd_config).
This is the next thing I tried after posting my previous reply.
It was commented, and the man page says it is enable by default.
Anyway, I forced it on yes, restarted sshd and tried again but with no
success.
I'm trying the same operation after increasing the sshd log verbosity.
OK, I'm progressing : the ssh issue seems to be gone.
I changed some parameters but not sure which one are relevant.
When comparing with a fresh install, here are the diffs :
root:/etc# diff /etc/ssh/sshd_config /tmp/sshd_config
43a44
LogLevel DEBUG3
48c49
< #PermitRootLogin yes
---
PermitRootLogin yes
114,115c115,116
< #AllowTcpForwarding yes
< #GatewayPorts no
---
AllowTcpForwarding yes
GatewayPorts yes
The man page specifically says that the -R option won't work unless
GatewayPorts is set to yes, and it is NOT able by default.
But i don't remember if this was sufficient, because I may have change
PermitRootLogin in the same time.
What is sure is that AllowTcpForwading set to yes did not change anything.
Anyway, now, I can see things running for some times. Then failing.
The log file is attached. I have read it quickly. I'm tired. See you
tomorrow.
Relevant parts :
--------------------------------------------
libguestfs: trace: hivex_node_get_child = 968144
libguestfs: trace: hivex_node_get_child 968144 "firstboot"
guestfsd: main_loop: new request, len 0x40^M
guestfsd: error: failed: Argument list too long^M
guestfsd: main_loop: proc 355 (hivex_node_get_child) took 0.00 seconds^M
libguestfs: trace: hivex_node_get_child = -1 (error)
virt-v2v: error: libguestfs error: hivex_node_get_child: failed: Argument
list too long
--------------------------------------------
--
Nicolas Ecarnot
4:28 a.m.
On Wed, Nov 19, 2014 at 11:47:11PM +0100, Nicolas Ecarnot wrote:
OK, I'm progressing : the ssh issue seems to be gone.
I changed some parameters but not sure which one are relevant.
When comparing with a fresh install, here are the diffs :
root:/etc# diff /etc/ssh/sshd_config /tmp/sshd_config
43a44
> LogLevel DEBUG3
48c49
< #PermitRootLogin yes
---
> PermitRootLogin yes
114,115c115,116
< #AllowTcpForwarding yes
< #GatewayPorts no
---
> AllowTcpForwarding yes
> GatewayPorts yes
The man page specifically says that the -R option won't work unless
GatewayPorts is set to yes, and it is NOT able by default.
But i don't remember if this was sufficient, because I may have
change PermitRootLogin in the same time.
My understanding is that GatewayPorts should not be necessary.
Does it still work if you set it back to 'no'?
What is sure is that AllowTcpForwading set to yes did not change
anything.
Anyway, now, I can see things running for some times. Then failing.
The log file is attached. I have read it quickly. I'm tired. See you
tomorrow.
Relevant parts :
--------------------------------------------
libguestfs: trace: hivex_node_get_child = 968144
libguestfs: trace: hivex_node_get_child 968144 "firstboot"
guestfsd: main_loop: new request, len 0x40^M
guestfsd: error: failed: Argument list too long^M
guestfsd: main_loop: proc 355 (hivex_node_get_child) took 0.00 seconds^M
libguestfs: trace: hivex_node_get_child = -1 (error)
virt-v2v: error: libguestfs error: hivex_node_get_child: failed: Argument
list too long
That would indicate some problem with the guest registry.
You could grab it and post it somewhere (but note that registries
contain sensitive data such as passwords):
mkdir /tmp/registry
virt-copy-out -a /path/to/guest.img 'win:C:\windows\system32\config'
/tmp/registry
(I only need to see the files 'SYSTEM' and 'SOFTWARE')
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
Fedora Windows cross-compiler. Compile Windows programs, test, and
build Windows installers. Over 100 libraries supported.
http://fedoraproject.org/wiki/MinGW
2:04 p.m.
Le 20/11/2014 11:28, Richard W.M. Jones a écrit :
On Wed, Nov 19, 2014 at 11:47:11PM +0100, Nicolas Ecarnot wrote:
> OK, I'm progressing : the ssh issue seems to be gone.
> I changed some parameters but not sure which one are relevant.
> When comparing with a fresh install, here are the diffs :
>
> root:/etc# diff /etc/ssh/sshd_config /tmp/sshd_config
> 43a44
>> LogLevel DEBUG3
> 48c49
> < #PermitRootLogin yes
> ---
>> PermitRootLogin yes
> 114,115c115,116
> < #AllowTcpForwarding yes
> < #GatewayPorts no
> ---
>> AllowTcpForwarding yes
>> GatewayPorts yes
>
> The man page specifically says that the -R option won't work unless
> GatewayPorts is set to yes, and it is NOT able by default.
> But i don't remember if this was sufficient, because I may have
> change PermitRootLogin in the same time.
My understanding is that GatewayPorts should not be necessary.
Does it still work if you set it back to 'no'?
Ok, the test I did just now with GatewayPorts no is still working.
You were right.
--
Nicolas Ecarnot
5:18 p.m.
On 11/19/2014 04:47 AM, Richard W.M. Jones wrote:
On Wed, Nov 19, 2014 at 09:08:36AM +0100, Nicolas Ecarnot wrote:
> I tried to use the virt-p2v-make-disk, but it failed to connect to a
> libvirt socket (obviously, there is none running in this VM.)
To avoid libvirt problems, do:
export LIBGUESTFS_BACKEND=direct
Why don't you use the ISO? Burn it onto a CD and boot it.
I've tried the
ISO,but I use usb disk.
# livecd-iso-to-disk /livecd-p2v-201411180515.iso /dev/sdb
However,when I use the usb disk to boot,it fails with the below error:
No default or UI configuration directive found.
Then it hangs with boot:.
Rich.
4:24 a.m.
On Wed, Nov 19, 2014 at 06:18:33PM -0500, tzheng wrote:
On 11/19/2014 04:47 AM, Richard W.M. Jones wrote:
>On Wed, Nov 19, 2014 at 09:08:36AM +0100, Nicolas Ecarnot wrote:
>>I tried to use the virt-p2v-make-disk, but it failed to connect to a
>>libvirt socket (obviously, there is none running in this VM.)
>To avoid libvirt problems, do:
>
> export LIBGUESTFS_BACKEND=direct
>
>Why don't you use the ISO? Burn it onto a CD and boot it.
I've tried the ISO,but I use usb disk.
# livecd-iso-to-disk /livecd-p2v-201411180515.iso /dev/sdb
However,when I use the usb disk to boot,it fails with the below error:
No default or UI configuration directive found.
Then it hangs with boot:.
This seems to be a bug in livecd-iso-to-disk. The whole of
livecd-tools seems to be buggy and now abandoned.
You can use virt-p2v-make-disk to make a bootable USB stick. It uses
virt-builder and so is a lot more reliable, and quicker.
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-df lists disk usage of guests without needing to install any
software inside the virtual machine. Supports Linux and Windows.
http://people.redhat.com/~rjones/virt-df/
3655
days inactive
3665
days old
28 comments
4 participants
participants (4)
-
Matthew Booth -
Nicolas Ecarnot -
Richard W.M. Jones -
tzheng