thr3ads.net - Xen users - [Xen-users] Xen a couple of questions [Jun 2007]

If this information is useful, please help other people find it:
Share via:

Octavian Teodorescu

2007-Jun-05 08:34 UTC

[Xen-users] Xen a couple of questions

Hi guys,

1. Regarding Centos and Fedora core 7 compared with fedora core 5. I''ve
seen that on fedora core 5 when you want to install xen you have to
install the following packages: xen, kernel-xen0 and kernel-xenU (of
course with the dependencies needed). But on Centos, FC7 and I think
redhat versions, you only have to install xen and kernel-xen, you don''t
have any kernel for the guest system. In my case I could only start a xen
guest (on FC7) with an older kernel-xenU installed from FC version 5.

My question is: Why does the newer releases of linux has xen kernel
prebuilt but just for dom0, not for the guest systems, and you can''t
even
find a domU kernel special for those systems?

2.My network topology in my home is like this:
--------
-router-
--------
   |
   |
   |
-----------         ------------
-linux xen-   ----  -other 2 pc-
-----------         ------------

The linux xen machine has two network interfaces and xen installed.
I want: -  one windows machine virtualized
        -  one linux machine for which I want to have a public ipaddress
(to put the ip in DMS on the router) and I want it to use eth0 (so
in this case the traffic can not be sniffed by other guest systems
or dom0).

ifconfig -a (on dom0) it shows like this:
eth0      Link encap:Ethernet  HWaddr 00:00:E8:76:E2:4D
          UP BROADCAST MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
          Interrupt:21 Base address:0x2000

eth1      Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
          inet addr:192.168.0.101  Bcast:192.168.0.255  Mask:255.255.255.0
          inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:198578 errors:0 dropped:0 overruns:0 frame:0
          TX packets:117290 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:267328989 (254.9 MiB)  TX bytes:8294632 (7.9 MiB)

lo        Link encap:Local Loopback
          inet addr:127.0.0.1  Mask:255.0.0.0
          inet6 addr: ::1/128 Scope:Host
          UP LOOPBACK RUNNING  MTU:16436  Metric:1
          RX packets:2689 errors:0 dropped:0 overruns:0 frame:0
          TX packets:2689 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:12510296 (11.9 MiB)  TX bytes:12510296 (11.9 MiB)

peth1     Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
          inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:198588 errors:0 dropped:0 overruns:0 frame:0
          TX packets:117311 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:100
          RX bytes:270906777 (258.3 MiB)  TX bytes:8813848 (8.4 MiB)
          Base address:0x40c0 Memory:92200000-92220000

vif4.0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
          inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:9 errors:0 dropped:0 overruns:0 frame:0
          TX packets:1 errors:0 dropped:6 overruns:0 carrier:0
          collisions:0 txqueuelen:1
          RX bytes:1068 (1.0 KiB)  TX bytes:342 (342.0 b)

virbr0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
          inet addr:192.168.122.1  Bcast:192.168.122.255  Mask:255.255.255.0
          inet6 addr: fe80::200:ff:fe00:0/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:43 errors:0 dropped:0 overruns:0 frame:0
          TX packets:17 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:3208 (3.1 KiB)  TX bytes:2018 (1.9 KiB)

I don''t see any xen bridge, because that''s what I think I
need: one
network card, and one xen bridge.
I found on google that I could use the following script:
#!/bin/sh
dir=$(dirname "$0")
"$dir/network-bridge" "$@" vifnum=0 netdev=eth0
bridge=xenbr0
"$dir/network-bridge" "$@" vifnum=1 netdev=eth1
bridge=xenbr1
"$dir/network-bridge" "$@" vifnum=2 netdev=eth2
bridge=xenbr2
And then set it into xen-config.sxp:
network-script matrix-network
But it gives errors that network-script has only start, stop and status.
The only thing that it succeds is that I can see a xen bridge. If this
would work, doesn''t this affects other guest domains also?

My question is: How can I set a guest dom to use directly a network card
with other ip class ?

Best regards,
Octav



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

Nico Kadel-Garcia

2007-Jun-05 08:56 UTC

head link

Re: [Xen-users] Xen a couple of questions

Octavian Teodorescu wrote:> Hi guys,
>
> 1. Regarding Centos and Fedora core 7 compared with fedora core 5.
I''ve
> seen that on fedora core 5 when you want to install xen you have to
> install the following packages: xen, kernel-xen0 and kernel-xenU (of
> course with the dependencies needed). But on Centos, FC7 and I think
> redhat versions, you only have to install xen and kernel-xen, you
don''t
> have any kernel for the guest system. In my case I could only start a xen
> guest (on FC7) with an older kernel-xenU installed from FC version 5.
>   Probably to ease kernel management. Having to build and test kernel-xen 
and kernel-xenU means two different sets of testing environments for 
other software: that''s asking for trouble in a big, commercial support 
environment like RedHat''s various enterprise customers, even if the
xenU
kernels can be made more efficient..

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

Geert Janssens

2007-Jun-05 09:04 UTC

head link

Re: [Xen-users] Xen a couple of questions

1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6 and up, the xen kernel 
can be used for both dom0 and domU. There is no need anymore for two kernels.

2. I don''t know the complete answer to your second question. From your 
ifconfig output, it looks as if Fedora Core 7 creates a virbr0 interface 
instead of a xenbr0 interface. You could check if this is really a bridge 
with the command "brctl show". You probably have to execute this as
root.

Then if I understand your question correctly, you are trying to setup a xen 
guest domain to act as a firewall/router/gateway/whatever for your lan.

So I assume you only want this guest domain to use the external network card 
(your eth0). There are two ways of accomplishing this:
* either use PCI passthrough so that your dom0 won''t see eth0, but
instead
it''s passed to your guest system (search for pciback on Google for more
info). Unfortunatly, I didn''t manage to set this up in my particular
case, so
I used the seconde option:
* create two xenbridges, one for your external network interface, and one for 
your internal network interface. Then configure dom0 such that it isn''t
allowed to use the bridge for the external interface. You can do this by 
either disabling the virtual interface in dom0 (which will be called eth0) or 
by setting some firewall rules in dom0, or both.
You can search this list for one of my earlier mails, where I explain my 
configuration (on CentOS 5). It''s titled "advanced
bridging..." and dated May
16th, 2007.

Hopefully this will help you along the way.

Cheers,

Geert

On Tuesday 5 June 2007 10:34, Octavian Teodorescu wrote:> Hi guys,
>
> 1. Regarding Centos and Fedora core 7 compared with fedora core 5.
I''ve
> seen that on fedora core 5 when you want to install xen you have to
> install the following packages: xen, kernel-xen0 and kernel-xenU (of
> course with the dependencies needed). But on Centos, FC7 and I think
> redhat versions, you only have to install xen and kernel-xen, you
don''t
> have any kernel for the guest system. In my case I could only start a xen
> guest (on FC7) with an older kernel-xenU installed from FC version 5.
>
> My question is: Why does the newer releases of linux has xen kernel
> prebuilt but just for dom0, not for the guest systems, and you
can''t even
> find a domU kernel special for those systems?
>
> 2.My network topology in my home is like this:
> --------
> -router-
> --------
>
>
>
> -----------         ------------
> -linux xen-   ----  -other 2 pc-
> -----------         ------------
>
> The linux xen machine has two network interfaces and xen installed.
> I want: -  one windows machine virtualized
>         -  one linux machine for which I want to have a public ipaddress
> (to put the ip in DMS on the router) and I want it to use eth0 (so
> in this case the traffic can not be sniffed by other guest systems
> or dom0).
>
> ifconfig -a (on dom0) it shows like this:
> eth0      Link encap:Ethernet  HWaddr 00:00:E8:76:E2:4D
>           UP BROADCAST MULTICAST  MTU:1500  Metric:1
>           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:1000
>           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
>           Interrupt:21 Base address:0x2000
>
> eth1      Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>           inet addr:192.168.0.101  Bcast:192.168.0.255  Mask:255.255.255.0
>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:198578 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:117290 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:267328989 (254.9 MiB)  TX bytes:8294632 (7.9 MiB)
>
> lo        Link encap:Local Loopback
>           inet addr:127.0.0.1  Mask:255.0.0.0
>           inet6 addr: ::1/128 Scope:Host
>           UP LOOPBACK RUNNING  MTU:16436  Metric:1
>           RX packets:2689 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:2689 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:12510296 (11.9 MiB)  TX bytes:12510296 (11.9 MiB)
>
> peth1     Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:198588 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:117311 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:100
>           RX bytes:270906777 (258.3 MiB)  TX bytes:8813848 (8.4 MiB)
>           Base address:0x40c0 Memory:92200000-92220000
>
> vif4.0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>           inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:9 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:1 errors:0 dropped:6 overruns:0 carrier:0
>           collisions:0 txqueuelen:1
>           RX bytes:1068 (1.0 KiB)  TX bytes:342 (342.0 b)
>
> virbr0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>           inet addr:192.168.122.1  Bcast:192.168.122.255 
> Mask:255.255.255.0 inet6 addr: fe80::200:ff:fe00:0/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:43 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:17 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:3208 (3.1 KiB)  TX bytes:2018 (1.9 KiB)
>
> I don''t see any xen bridge, because that''s what I think I
need: one
> network card, and one xen bridge.
> I found on google that I could use the following script:
> #!/bin/sh
> dir=$(dirname "$0")
> "$dir/network-bridge" "$@" vifnum=0 netdev=eth0
bridge=xenbr0
> "$dir/network-bridge" "$@" vifnum=1 netdev=eth1
bridge=xenbr1
> "$dir/network-bridge" "$@" vifnum=2 netdev=eth2
bridge=xenbr2
> And then set it into xen-config.sxp:
> network-script matrix-network
> But it gives errors that network-script has only start, stop and status.
> The only thing that it succeds is that I can see a xen bridge. If this
> would work, doesn''t this affects other guest domains also?
>
> My question is: How can I set a guest dom to use directly a network card
> with other ip class ?
>
> Best regards,
> Octav
>
>
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@lists.xensource.com
> http://lists.xensource.com/xen-users
-- 
Kobalt W.I.T.
Web & Information Technology
Brusselsesteenweg 152
1850 Grimbergen

Tel  : +32 479 339 655
Email: info@kobaltwit.be

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

Nico Kadel-Garcia

2007-Jun-05 09:26 UTC

head link

RHEL 5 glibc, was Re: [Xen-users] Xen a couple of questions

Geert Janssens wrote:> 1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6 and up, the xen
kernel
> can be used for both dom0 and domU. There is no need anymore for two
kernels.
>   Has anyone properly tested the RHEL and CentOS 5 glibc in clients? I''m 
forced to use the xensource 3.1.0 kernel-xen kernel to get RHEL 4 
DomU''s, and it causes the guests to whine bitterly at boot time about 
the /lib/tls issues. Xensource used to publish a modified glibc for RHEL 
4.4 to avoid the /lib/tls emulation issues, but CentOS and RHEL 4.5 have 
an even more recent glibc.



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

Michael Paesold

2007-Jun-05 10:06 UTC

head link

Re: RHEL 5 glibc, was Re: [Xen-users] Xen a couple of questions

Nico Kadel-Garcia wrote:> Geert Janssens wrote:
>> 1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6 and up, the xen 
>> kernel can be used for both dom0 and domU. There is no need anymore 
>> for two kernels.
>>   
> Has anyone properly tested the RHEL and CentOS 5 glibc in clients?
I''m
> forced to use the xensource 3.1.0 kernel-xen kernel to get RHEL 4 
> DomU''s, and it causes the guests to whine bitterly at boot time
about
> the /lib/tls issues. Xensource used to publish a modified glibc for RHEL 
> 4.4 to avoid the /lib/tls emulation issues, but CentOS and RHEL 4.5 have 
> an even more recent glibc.
What kernel do you use in the guests? The RedHat/CentOS provided kernels 
properly place glibc hint files into /etc/ld.so.conf.d/, which makes 
glibc use the nosegneg variants available in RHEL/CentOS 4.5 and 5.

If you use Xensource kernels, I guess those don''t create the files 
correctly.

Best Regards
Michael Paesold


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

Octavian Teodorescu

2007-Jun-05 11:11 UTC

head link

Re: [Xen-users] Xen a couple of questions

Thanks a lot, I appreciate your help.
1. I tried with the same configuration and with the kernel of dom0, but I
receive a lot of errors, on both Fedora Core 7 and Centos 5 (systems with
which I''ve tried xen).

2.Yap that''s a bridge interface. I''ll look into your mails
about advance
bridging, thanks.

Best regards.
>> 1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6 and up, the xen
>> kernel
>> can be used for both dom0 and domU. There is no need anymore for two
>> kernels.
>>
>> 2. I don''t know the complete answer to your second question.
From your
>> ifconfig output, it looks as if Fedora Core 7 creates a virbr0
interface
>> instead of a xenbr0 interface. You could check if this is really a
bridge
>> with the command "brctl show". You probably have to execute
this as root.
>>
>> Then if I understand your question correctly, you are trying to setup a
>> xen
>> guest domain to act as a firewall/router/gateway/whatever for your lan.
>>
>> So I assume you only want this guest domain to use the external network
>> card
>> (your eth0). There are two ways of accomplishing this:
>> * either use PCI passthrough so that your dom0 won''t see eth0,
but instead
>> it''s passed to your guest system (search for pciback on Google
for more
>> info). Unfortunatly, I didn''t manage to set this up in my
particular case,
>> so
>> I used the seconde option:
>> * create two xenbridges, one for your external network interface, and
one
>> for
>> your internal network interface. Then configure dom0 such that it
isn''t
>> allowed to use the bridge for the external interface. You can do this
by
>> either disabling the virtual interface in dom0 (which will be called
eth0)
>> or
>> by setting some firewall rules in dom0, or both.
>> You can search this list for one of my earlier mails, where I explain
my
>> configuration (on CentOS 5). It''s titled "advanced
bridging..." and dated
>> May
>> 16th, 2007.
>>
>> Hopefully this will help you along the way.
>>
>> Cheers,
>>
>> Geert
>>
>> On Tuesday 5 June 2007 10:34, Octavian Teodorescu wrote:
>>> Hi guys,
>>>
>>> 1. Regarding Centos and Fedora core 7 compared with fedora core 5.
I''ve
>>> seen that on fedora core 5 when you want to install xen you have to
>>> install the following packages: xen, kernel-xen0 and kernel-xenU
(of
>>> course with the dependencies needed). But on Centos, FC7 and I
think
>>> redhat versions, you only have to install xen and kernel-xen, you
don''t
>>> have any kernel for the guest system. In my case I could only start
a
>>> xen
>>> guest (on FC7) with an older kernel-xenU installed from FC version
5.
>>>
>>> My question is: Why does the newer releases of linux has xen kernel
>>> prebuilt but just for dom0, not for the guest systems, and you
can''t
>>> even
>>> find a domU kernel special for those systems?
>>>
>>> 2.My network topology in my home is like this:
>>> --------
>>> -router-
>>> --------
>>>
>>>
>>>
>>> -----------         ------------
>>> -linux xen-   ----  -other 2 pc-
>>> -----------         ------------
>>>
>>> The linux xen machine has two network interfaces and xen installed.
>>> I want: -  one windows machine virtualized
>>>         -  one linux machine for which I want to have a public
ipaddress
>>> (to put the ip in DMS on the router) and I want it to use eth0 (so
>>> in this case the traffic can not be sniffed by other guest systems
>>> or dom0).
>>>
>>> ifconfig -a (on dom0) it shows like this:
>>> eth0      Link encap:Ethernet  HWaddr 00:00:E8:76:E2:4D
>>>           UP BROADCAST MULTICAST  MTU:1500  Metric:1
>>>           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
>>>           TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
>>>           collisions:0 txqueuelen:1000
>>>           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
>>>           Interrupt:21 Base address:0x2000
>>>
>>> eth1      Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>>>           inet addr:192.168.0.101  Bcast:192.168.0.255
>>> Mask:255.255.255.0
>>>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
>>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>>>           RX packets:198578 errors:0 dropped:0 overruns:0 frame:0
>>>           TX packets:117290 errors:0 dropped:0 overruns:0 carrier:0
>>>           collisions:0 txqueuelen:0
>>>           RX bytes:267328989 (254.9 MiB)  TX bytes:8294632 (7.9
MiB)
>>>
>>> lo        Link encap:Local Loopback
>>>           inet addr:127.0.0.1  Mask:255.0.0.0
>>>           inet6 addr: ::1/128 Scope:Host
>>>           UP LOOPBACK RUNNING  MTU:16436  Metric:1
>>>           RX packets:2689 errors:0 dropped:0 overruns:0 frame:0
>>>           TX packets:2689 errors:0 dropped:0 overruns:0 carrier:0
>>>           collisions:0 txqueuelen:0
>>>           RX bytes:12510296 (11.9 MiB)  TX bytes:12510296 (11.9
MiB)
>>>
>>> peth1     Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>>>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
>>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>>>           RX packets:198588 errors:0 dropped:0 overruns:0 frame:0
>>>           TX packets:117311 errors:0 dropped:0 overruns:0 carrier:0
>>>           collisions:0 txqueuelen:100
>>>           RX bytes:270906777 (258.3 MiB)  TX bytes:8813848 (8.4
MiB)
>>>           Base address:0x40c0 Memory:92200000-92220000
>>>
>>> vif4.0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>>>           inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
>>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>>>           RX packets:9 errors:0 dropped:0 overruns:0 frame:0
>>>           TX packets:1 errors:0 dropped:6 overruns:0 carrier:0
>>>           collisions:0 txqueuelen:1
>>>           RX bytes:1068 (1.0 KiB)  TX bytes:342 (342.0 b)
>>>
>>> virbr0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>>>           inet addr:192.168.122.1  Bcast:192.168.122.255
>>> Mask:255.255.255.0 inet6 addr: fe80::200:ff:fe00:0/64 Scope:Link
>>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>>>           RX packets:43 errors:0 dropped:0 overruns:0 frame:0
>>>           TX packets:17 errors:0 dropped:0 overruns:0 carrier:0
>>>           collisions:0 txqueuelen:0
>>>           RX bytes:3208 (3.1 KiB)  TX bytes:2018 (1.9 KiB)
>>>
>>> I don''t see any xen bridge, because that''s what I
think I need: one
>>> network card, and one xen bridge.
>>> I found on google that I could use the following script:
>>> #!/bin/sh
>>> dir=$(dirname "$0")
>>> "$dir/network-bridge" "$@" vifnum=0 netdev=eth0
bridge=xenbr0
>>> "$dir/network-bridge" "$@" vifnum=1 netdev=eth1
bridge=xenbr1
>>> "$dir/network-bridge" "$@" vifnum=2 netdev=eth2
bridge=xenbr2
>>> And then set it into xen-config.sxp:
>>> network-script matrix-network
>>> But it gives errors that network-script has only start, stop and
status.
>>> The only thing that it succeds is that I can see a xen bridge. If
this
>>> would work, doesn''t this affects other guest domains also?
>>>
>>> My question is: How can I set a guest dom to use directly a network
card
>>> with other ip class ?
>>>
>>> Best regards,
>>> Octav
>>>
>>>
>>>
>>> _______________________________________________
>>> Xen-users mailing list
>>> Xen-users@lists.xensource.com
>>> http://lists.xensource.com/xen-users
>>
>> --
>> Kobalt W.I.T.
>> Web & Information Technology
>> Brusselsesteenweg 152
>> 1850 Grimbergen
>>
>> Tel  : +32 479 339 655
>> Email: info@kobaltwit.be
>>
>> _______________________________________________
>> Xen-users mailing list
>> Xen-users@lists.xensource.com
>> http://lists.xensource.com/xen-users
>>


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

Geert Janssens

2007-Jun-05 12:05 UTC

head link

Re: [Xen-users] Xen a couple of questions

On Tuesday 5 June 2007 13:11, you wrote:> Thanks a lot, I appreciate your help.
> 1. I tried with the same configuration and with the kernel of dom0, but I
> receive a lot of errors, on both Fedora Core 7 and Centos 5 (systems with
> which I''ve tried xen).
>Hmm, I don''t know what exactly you tried and what is failing.

I simply used virt-manager to create a CentOS 5 guest on my CentOS 5 dom0. I 
followed (more or less) the guidelines that come with CentOS 5''s
release:
http://www.centos.org/docs/5/html/Virtualization-en-US/

I don''t remember having particular difficulty with this. The only
caveat I
remember was that Anaconda insists on a block device that can hold a 
partition map. So while installing, you can''t provide the guest with
separate
partitions, because it will treat these separate partitions as complete 
diskss that have to be partitioned still.

Regards,

Geert
> 2.Yap that''s a bridge interface. I''ll look into your
mails about advance
> bridging, thanks.
>
> Best regards.
>
> >> 1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6 and up, the
xen
> >> kernel
> >> can be used for both dom0 and domU. There is no need anymore for
two
> >> kernels.
> >>
> >> 2. I don''t know the complete answer to your second
question. From your
> >> ifconfig output, it looks as if Fedora Core 7 creates a virbr0
interface
> >> instead of a xenbr0 interface. You could check if this is really a
> >> bridge with the command "brctl show". You probably have
to execute this
> >> as root.
> >>
> >> Then if I understand your question correctly, you are trying to
setup a
> >> xen
> >> guest domain to act as a firewall/router/gateway/whatever for your
lan.
> >>
> >> So I assume you only want this guest domain to use the external
network
> >> card
> >> (your eth0). There are two ways of accomplishing this:
> >> * either use PCI passthrough so that your dom0 won''t see
eth0, but
> >> instead it''s passed to your guest system (search for
pciback on Google
> >> for more info). Unfortunatly, I didn''t manage to set this
up in my
> >> particular case, so
> >> I used the seconde option:
> >> * create two xenbridges, one for your external network interface,
and
> >> one for
> >> your internal network interface. Then configure dom0 such that it
isn''t
> >> allowed to use the bridge for the external interface. You can do
this by
> >> either disabling the virtual interface in dom0 (which will be
called
> >> eth0) or
> >> by setting some firewall rules in dom0, or both.
> >> You can search this list for one of my earlier mails, where I
explain my
> >> configuration (on CentOS 5). It''s titled "advanced
bridging..." and
> >> dated May
> >> 16th, 2007.
> >>
> >> Hopefully this will help you along the way.
> >>
> >> Cheers,
> >>
> >> Geert
> >>
> >> On Tuesday 5 June 2007 10:34, Octavian Teodorescu wrote:
> >>> Hi guys,
> >>>
> >>> 1. Regarding Centos and Fedora core 7 compared with fedora
core 5. I''ve
> >>> seen that on fedora core 5 when you want to install xen you
have to
> >>> install the following packages: xen, kernel-xen0 and
kernel-xenU (of
> >>> course with the dependencies needed). But on Centos, FC7 and I
think
> >>> redhat versions, you only have to install xen and kernel-xen,
you don''t
> >>> have any kernel for the guest system. In my case I could only
start a
> >>> xen
> >>> guest (on FC7) with an older kernel-xenU installed from FC
version 5.
> >>>
> >>> My question is: Why does the newer releases of linux has xen
kernel
> >>> prebuilt but just for dom0, not for the guest systems, and you
can''t
> >>> even
> >>> find a domU kernel special for those systems?
> >>>
> >>> 2.My network topology in my home is like this:
> >>> --------
> >>> -router-
> >>> --------
> >>>
> >>>
> >>>
> >>> -----------         ------------
> >>> -linux xen-   ----  -other 2 pc-
> >>> -----------         ------------
> >>>
> >>> The linux xen machine has two network interfaces and xen
installed.
> >>> I want: -  one windows machine virtualized
> >>>         -  one linux machine for which I want to have a public
> >>> ipaddress (to put the ip in DMS on the router) and I want it
to use
> >>> eth0 (so in this case the traffic can not be sniffed by other
guest
> >>> systems or dom0).
> >>>
> >>> ifconfig -a (on dom0) it shows like this:
> >>> eth0      Link encap:Ethernet  HWaddr 00:00:E8:76:E2:4D
> >>>           UP BROADCAST MULTICAST  MTU:1500  Metric:1
> >>>           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
> >>>           TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
> >>>           collisions:0 txqueuelen:1000
> >>>           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
> >>>           Interrupt:21 Base address:0x2000
> >>>
> >>> eth1      Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
> >>>           inet addr:192.168.0.101  Bcast:192.168.0.255
> >>> Mask:255.255.255.0
> >>>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
> >>>           RX packets:198578 errors:0 dropped:0 overruns:0
frame:0
> >>>           TX packets:117290 errors:0 dropped:0 overruns:0
carrier:0
> >>>           collisions:0 txqueuelen:0
> >>>           RX bytes:267328989 (254.9 MiB)  TX bytes:8294632
(7.9 MiB)
> >>>
> >>> lo        Link encap:Local Loopback
> >>>           inet addr:127.0.0.1  Mask:255.0.0.0
> >>>           inet6 addr: ::1/128 Scope:Host
> >>>           UP LOOPBACK RUNNING  MTU:16436  Metric:1
> >>>           RX packets:2689 errors:0 dropped:0 overruns:0
frame:0
> >>>           TX packets:2689 errors:0 dropped:0 overruns:0
carrier:0
> >>>           collisions:0 txqueuelen:0
> >>>           RX bytes:12510296 (11.9 MiB)  TX bytes:12510296
(11.9 MiB)
> >>>
> >>> peth1     Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
> >>>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
> >>>           RX packets:198588 errors:0 dropped:0 overruns:0
frame:0
> >>>           TX packets:117311 errors:0 dropped:0 overruns:0
carrier:0
> >>>           collisions:0 txqueuelen:100
> >>>           RX bytes:270906777 (258.3 MiB)  TX bytes:8813848
(8.4 MiB)
> >>>           Base address:0x40c0 Memory:92200000-92220000
> >>>
> >>> vif4.0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
> >>>           inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
> >>>           RX packets:9 errors:0 dropped:0 overruns:0 frame:0
> >>>           TX packets:1 errors:0 dropped:6 overruns:0 carrier:0
> >>>           collisions:0 txqueuelen:1
> >>>           RX bytes:1068 (1.0 KiB)  TX bytes:342 (342.0 b)
> >>>
> >>> virbr0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
> >>>           inet addr:192.168.122.1  Bcast:192.168.122.255
> >>> Mask:255.255.255.0 inet6 addr: fe80::200:ff:fe00:0/64
Scope:Link
> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
> >>>           RX packets:43 errors:0 dropped:0 overruns:0 frame:0
> >>>           TX packets:17 errors:0 dropped:0 overruns:0
carrier:0
> >>>           collisions:0 txqueuelen:0
> >>>           RX bytes:3208 (3.1 KiB)  TX bytes:2018 (1.9 KiB)
> >>>
> >>> I don''t see any xen bridge, because that''s
what I think I need: one
> >>> network card, and one xen bridge.
> >>> I found on google that I could use the following script:
> >>> #!/bin/sh
> >>> dir=$(dirname "$0")
> >>> "$dir/network-bridge" "$@" vifnum=0
netdev=eth0 bridge=xenbr0
> >>> "$dir/network-bridge" "$@" vifnum=1
netdev=eth1 bridge=xenbr1
> >>> "$dir/network-bridge" "$@" vifnum=2
netdev=eth2 bridge=xenbr2
> >>> And then set it into xen-config.sxp:
> >>> network-script matrix-network
> >>> But it gives errors that network-script has only start, stop
and
> >>> status. The only thing that it succeds is that I can see a xen
bridge.
> >>> If this would work, doesn''t this affects other guest
domains also?
> >>>
> >>> My question is: How can I set a guest dom to use directly a
network
> >>> card with other ip class ?
> >>>
> >>> Best regards,
> >>> Octav
> >>>
> >>>
> >>>
> >>> _______________________________________________
> >>> Xen-users mailing list
> >>> Xen-users@lists.xensource.com
> >>> http://lists.xensource.com/xen-users
> >>
> >> --
> >> Kobalt W.I.T.
> >> Web & Information Technology
> >> Brusselsesteenweg 152
> >> 1850 Grimbergen
> >>
> >> Tel  : +32 479 339 655
> >> Email: info@kobaltwit.be
> >>
> >> _______________________________________________
> >> Xen-users mailing list
> >> Xen-users@lists.xensource.com
> >> http://lists.xensource.com/xen-users
-- 
Kobalt W.I.T.
Web & Information Technology
Brusselsesteenweg 152
1850 Grimbergen

Tel  : +32 479 339 655
Email: info@kobaltwit.be

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

Octavian Teodorescu

2007-Jun-05 12:19 UTC

head link

Re: [Xen-users] Xen a couple of questions

I had stoped the virtual domain and modified in the config file to use the
dom0 kernel.
Here are the last lines of the guest boot where you can see the error:
"SCSI subsystem initialized
device-mapper: ioctl: 4.11.0-ioctl (2006-10-12) initialised:
dm-devel@redhat.com
Kernel panic - not syncing: Attempted to kill init!"

But if I use the old kernelU i have which is actually for Fedorac core 5
then everthing it''s ok.

>> On Tuesday 5 June 2007 13:11, you wrote:
>>> Thanks a lot, I appreciate your help.
>>> 1. I tried with the same configuration and with the kernel of dom0,
but
>>> I
>>> receive a lot of errors, on both Fedora Core 7 and Centos 5
(systems
>>> with
>>> which I''ve tried xen).
>>>
>> Hmm, I don''t know what exactly you tried and what is failing.
>>
>> I simply used virt-manager to create a CentOS 5 guest on my CentOS 5
dom0.
>> I
>> followed (more or less) the guidelines that come with CentOS
5''s release:
>> http://www.centos.org/docs/5/html/Virtualization-en-US/
>>
>> I don''t remember having particular difficulty with this. The
only caveat I
>> remember was that Anaconda insists on a block device that can hold a
>> partition map. So while installing, you can''t provide the
guest with
>> separate
>> partitions, because it will treat these separate partitions as complete
>> diskss that have to be partitioned still.
>>
>> Regards,
>>
>> Geert
>>
>>> 2.Yap that''s a bridge interface. I''ll look into
your mails about advance
>>> bridging, thanks.
>>>
>>> Best regards.
>>>
>>> >> 1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6 and
up, the xen
>>> >> kernel
>>> >> can be used for both dom0 and domU. There is no need
anymore for two
>>> >> kernels.
>>> >>
>>> >> 2. I don''t know the complete answer to your
second question. From
>>> your
>>> >> ifconfig output, it looks as if Fedora Core 7 creates a
virbr0
>>> interface
>>> >> instead of a xenbr0 interface. You could check if this is
really a
>>> >> bridge with the command "brctl show". You
probably have to execute
>>> this
>>> >> as root.
>>> >>
>>> >> Then if I understand your question correctly, you are
trying to setup
>>> a
>>> >> xen
>>> >> guest domain to act as a firewall/router/gateway/whatever
for your
>>> lan.
>>> >>
>>> >> So I assume you only want this guest domain to use the
external
>>> network
>>> >> card
>>> >> (your eth0). There are two ways of accomplishing this:
>>> >> * either use PCI passthrough so that your dom0
won''t see eth0, but
>>> >> instead it''s passed to your guest system (search
for pciback on
>>> Google
>>> >> for more info). Unfortunatly, I didn''t manage to
set this up in my
>>> >> particular case, so
>>> >> I used the seconde option:
>>> >> * create two xenbridges, one for your external network
interface, and
>>> >> one for
>>> >> your internal network interface. Then configure dom0 such
that it
>>> isn''t
>>> >> allowed to use the bridge for the external interface. You
can do this
>>> by
>>> >> either disabling the virtual interface in dom0 (which will
be called
>>> >> eth0) or
>>> >> by setting some firewall rules in dom0, or both.
>>> >> You can search this list for one of my earlier mails,
where I explain
>>> my
>>> >> configuration (on CentOS 5). It''s titled
"advanced bridging..." and
>>> >> dated May
>>> >> 16th, 2007.
>>> >>
>>> >> Hopefully this will help you along the way.
>>> >>
>>> >> Cheers,
>>> >>
>>> >> Geert
>>> >>
>>> >> On Tuesday 5 June 2007 10:34, Octavian Teodorescu wrote:
>>> >>> Hi guys,
>>> >>>
>>> >>> 1. Regarding Centos and Fedora core 7 compared with
fedora core 5.
>>> I''ve
>>> >>> seen that on fedora core 5 when you want to install
xen you have to
>>> >>> install the following packages: xen, kernel-xen0 and
kernel-xenU (of
>>> >>> course with the dependencies needed). But on Centos,
FC7 and I think
>>> >>> redhat versions, you only have to install xen and
kernel-xen, you
>>> don''t
>>> >>> have any kernel for the guest system. In my case I
could only start
>>> a
>>> >>> xen
>>> >>> guest (on FC7) with an older kernel-xenU installed
from FC version
>>> 5.
>>> >>>
>>> >>> My question is: Why does the newer releases of linux
has xen kernel
>>> >>> prebuilt but just for dom0, not for the guest systems,
and you can''t
>>> >>> even
>>> >>> find a domU kernel special for those systems?
>>> >>>
>>> >>> 2.My network topology in my home is like this:
>>> >>> --------
>>> >>> -router-
>>> >>> --------
>>> >>>
>>> >>>
>>> >>>
>>> >>> -----------         ------------
>>> >>> -linux xen-   ----  -other 2 pc-
>>> >>> -----------         ------------
>>> >>>
>>> >>> The linux xen machine has two network interfaces and
xen installed.
>>> >>> I want: -  one windows machine virtualized
>>> >>>         -  one linux machine for which I want to have
a public
>>> >>> ipaddress (to put the ip in DMS on the router) and I
want it to use
>>> >>> eth0 (so in this case the traffic can not be sniffed
by other guest
>>> >>> systems or dom0).
>>> >>>
>>> >>> ifconfig -a (on dom0) it shows like this:
>>> >>> eth0      Link encap:Ethernet  HWaddr
00:00:E8:76:E2:4D
>>> >>>           UP BROADCAST MULTICAST  MTU:1500  Metric:1
>>> >>>           RX packets:0 errors:0 dropped:0 overruns:0
frame:0
>>> >>>           TX packets:0 errors:0 dropped:0 overruns:0
carrier:0
>>> >>>           collisions:0 txqueuelen:1000
>>> >>>           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
>>> >>>           Interrupt:21 Base address:0x2000
>>> >>>
>>> >>> eth1      Link encap:Ethernet  HWaddr
00:16:76:B3:16:AB
>>> >>>           inet addr:192.168.0.101  Bcast:192.168.0.255
>>> >>> Mask:255.255.255.0
>>> >>>           inet6 addr: fe80::216:76ff:feb3:16ab/64
Scope:Link
>>> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500 
Metric:1
>>> >>>           RX packets:198578 errors:0 dropped:0
overruns:0 frame:0
>>> >>>           TX packets:117290 errors:0 dropped:0
overruns:0 carrier:0
>>> >>>           collisions:0 txqueuelen:0
>>> >>>           RX bytes:267328989 (254.9 MiB)  TX
bytes:8294632 (7.9 MiB)
>>> >>>
>>> >>> lo        Link encap:Local Loopback
>>> >>>           inet addr:127.0.0.1  Mask:255.0.0.0
>>> >>>           inet6 addr: ::1/128 Scope:Host
>>> >>>           UP LOOPBACK RUNNING  MTU:16436  Metric:1
>>> >>>           RX packets:2689 errors:0 dropped:0
overruns:0 frame:0
>>> >>>           TX packets:2689 errors:0 dropped:0
overruns:0 carrier:0
>>> >>>           collisions:0 txqueuelen:0
>>> >>>           RX bytes:12510296 (11.9 MiB)  TX
bytes:12510296 (11.9 MiB)
>>> >>>
>>> >>> peth1     Link encap:Ethernet  HWaddr
00:16:76:B3:16:AB
>>> >>>           inet6 addr: fe80::216:76ff:feb3:16ab/64
Scope:Link
>>> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500 
Metric:1
>>> >>>           RX packets:198588 errors:0 dropped:0
overruns:0 frame:0
>>> >>>           TX packets:117311 errors:0 dropped:0
overruns:0 carrier:0
>>> >>>           collisions:0 txqueuelen:100
>>> >>>           RX bytes:270906777 (258.3 MiB)  TX
bytes:8813848 (8.4 MiB)
>>> >>>           Base address:0x40c0 Memory:92200000-92220000
>>> >>>
>>> >>> vif4.0    Link encap:Ethernet  HWaddr
FE:FF:FF:FF:FF:FF
>>> >>>           inet6 addr: fe80::fcff:ffff:feff:ffff/64
Scope:Link
>>> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500 
Metric:1
>>> >>>           RX packets:9 errors:0 dropped:0 overruns:0
frame:0
>>> >>>           TX packets:1 errors:0 dropped:6 overruns:0
carrier:0
>>> >>>           collisions:0 txqueuelen:1
>>> >>>           RX bytes:1068 (1.0 KiB)  TX bytes:342 (342.0
b)
>>> >>>
>>> >>> virbr0    Link encap:Ethernet  HWaddr
FE:FF:FF:FF:FF:FF
>>> >>>           inet addr:192.168.122.1 
Bcast:192.168.122.255
>>> >>> Mask:255.255.255.0 inet6 addr: fe80::200:ff:fe00:0/64
Scope:Link
>>> >>>           UP BROADCAST RUNNING MULTICAST  MTU:1500 
Metric:1
>>> >>>           RX packets:43 errors:0 dropped:0 overruns:0
frame:0
>>> >>>           TX packets:17 errors:0 dropped:0 overruns:0
carrier:0
>>> >>>           collisions:0 txqueuelen:0
>>> >>>           RX bytes:3208 (3.1 KiB)  TX bytes:2018 (1.9
KiB)
>>> >>>
>>> >>> I don''t see any xen bridge, because
that''s what I think I need: one
>>> >>> network card, and one xen bridge.
>>> >>> I found on google that I could use the following
script:
>>> >>> #!/bin/sh
>>> >>> dir=$(dirname "$0")
>>> >>> "$dir/network-bridge" "$@"
vifnum=0 netdev=eth0 bridge=xenbr0
>>> >>> "$dir/network-bridge" "$@"
vifnum=1 netdev=eth1 bridge=xenbr1
>>> >>> "$dir/network-bridge" "$@"
vifnum=2 netdev=eth2 bridge=xenbr2
>>> >>> And then set it into xen-config.sxp:
>>> >>> network-script matrix-network
>>> >>> But it gives errors that network-script has only
start, stop and
>>> >>> status. The only thing that it succeds is that I can
see a xen
>>> bridge.
>>> >>> If this would work, doesn''t this affects
other guest domains also?
>>> >>>
>>> >>> My question is: How can I set a guest dom to use
directly a network
>>> >>> card with other ip class ?
>>> >>>
>>> >>> Best regards,
>>> >>> Octav
>>> >>>
>>> >>>
>>> >>>
>>> >>> _______________________________________________
>>> >>> Xen-users mailing list
>>> >>> Xen-users@lists.xensource.com
>>> >>> http://lists.xensource.com/xen-users
>>> >>
>>> >> --
>>> >> Kobalt W.I.T.
>>> >> Web & Information Technology
>>> >> Brusselsesteenweg 152
>>> >> 1850 Grimbergen
>>> >>
>>> >> Tel  : +32 479 339 655
>>> >> Email: info@kobaltwit.be
>>> >>
>>> >> _______________________________________________
>>> >> Xen-users mailing list
>>> >> Xen-users@lists.xensource.com
>>> >> http://lists.xensource.com/xen-users
>>
>> --
>> Kobalt W.I.T.
>> Web & Information Technology
>> Brusselsesteenweg 152
>> 1850 Grimbergen
>>
>> Tel  : +32 479 339 655
>> Email: info@kobaltwit.be
>>
>> _______________________________________________
>> Xen-users mailing list
>> Xen-users@lists.xensource.com
>> http://lists.xensource.com/xen-users
>>


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

Geert Janssens

2007-Jun-05 12:26 UTC

head link

Re: [Xen-users] Xen a couple of questions

I am sorry, this is way over my head. I''m just a (relatively new) user
of xen
myself. I hope someone else can help you here.

Regards,

Geert

On Tuesday 5 June 2007 14:19, Octavian Teodorescu wrote:> I had stoped the virtual domain and modified in the config file to use the
> dom0 kernel.
> Here are the last lines of the guest boot where you can see the error:
> "SCSI subsystem initialized
> device-mapper: ioctl: 4.11.0-ioctl (2006-10-12) initialised:
> dm-devel@redhat.com
> Kernel panic - not syncing: Attempted to kill init!"
>
> But if I use the old kernelU i have which is actually for Fedorac core 5
> then everthing it''s ok.
>
> >> On Tuesday 5 June 2007 13:11, you wrote:
> >>> Thanks a lot, I appreciate your help.
> >>> 1. I tried with the same configuration and with the kernel of
dom0, but
> >>> I
> >>> receive a lot of errors, on both Fedora Core 7 and Centos 5
(systems
> >>> with
> >>> which I''ve tried xen).
> >>
> >> Hmm, I don''t know what exactly you tried and what is
failing.
> >>
> >> I simply used virt-manager to create a CentOS 5 guest on my CentOS
5
> >> dom0. I
> >> followed (more or less) the guidelines that come with CentOS
5''s
> >> release: http://www.centos.org/docs/5/html/Virtualization-en-US/
> >>
> >> I don''t remember having particular difficulty with this.
The only caveat
> >> I remember was that Anaconda insists on a block device that can
hold a
> >> partition map. So while installing, you can''t provide the
guest with
> >> separate
> >> partitions, because it will treat these separate partitions as
complete
> >> diskss that have to be partitioned still.
> >>
> >> Regards,
> >>
> >> Geert
> >>
> >>> 2.Yap that''s a bridge interface. I''ll look
into your mails about
> >>> advance bridging, thanks.
> >>>
> >>> Best regards.
> >>>
> >>> >> 1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6
and up, the
> >>> >> xen kernel
> >>> >> can be used for both dom0 and domU. There is no need
anymore for two
> >>> >> kernels.
> >>> >>
> >>> >> 2. I don''t know the complete answer to your
second question. From
> >>>
> >>> your
> >>>
> >>> >> ifconfig output, it looks as if Fedora Core 7 creates
a virbr0
> >>>
> >>> interface
> >>>
> >>> >> instead of a xenbr0 interface. You could check if
this is really a
> >>> >> bridge with the command "brctl show". You
probably have to execute
> >>>
> >>> this
> >>>
> >>> >> as root.
> >>> >>
> >>> >> Then if I understand your question correctly, you are
trying to
> >>> >> setup
> >>>
> >>> a
> >>>
> >>> >> xen
> >>> >> guest domain to act as a
firewall/router/gateway/whatever for your
> >>>
> >>> lan.
> >>>
> >>> >> So I assume you only want this guest domain to use
the external
> >>>
> >>> network
> >>>
> >>> >> card
> >>> >> (your eth0). There are two ways of accomplishing
this:
> >>> >> * either use PCI passthrough so that your dom0
won''t see eth0, but
> >>> >> instead it''s passed to your guest system
(search for pciback on
> >>>
> >>> Google
> >>>
> >>> >> for more info). Unfortunatly, I didn''t
manage to set this up in my
> >>> >> particular case, so
> >>> >> I used the seconde option:
> >>> >> * create two xenbridges, one for your external
network interface,
> >>> >> and one for
> >>> >> your internal network interface. Then configure dom0
such that it
> >>>
> >>> isn''t
> >>>
> >>> >> allowed to use the bridge for the external interface.
You can do
> >>> >> this
> >>>
> >>> by
> >>>
> >>> >> either disabling the virtual interface in dom0 (which
will be called
> >>> >> eth0) or
> >>> >> by setting some firewall rules in dom0, or both.
> >>> >> You can search this list for one of my earlier mails,
where I
> >>> >> explain
> >>>
> >>> my
> >>>
> >>> >> configuration (on CentOS 5). It''s titled
"advanced bridging..." and
> >>> >> dated May
> >>> >> 16th, 2007.
> >>> >>
> >>> >> Hopefully this will help you along the way.
> >>> >>
> >>> >> Cheers,
> >>> >>
> >>> >> Geert
> >>> >>
> >>> >> On Tuesday 5 June 2007 10:34, Octavian Teodorescu
wrote:
> >>> >>> Hi guys,
> >>> >>>
> >>> >>> 1. Regarding Centos and Fedora core 7 compared
with fedora core 5.
> >>>
> >>> I''ve
> >>>
> >>> >>> seen that on fedora core 5 when you want to
install xen you have to
> >>> >>> install the following packages: xen, kernel-xen0
and kernel-xenU
> >>> >>> (of course with the dependencies needed). But on
Centos, FC7 and I
> >>> >>> think redhat versions, you only have to install
xen and kernel-xen,
> >>> >>> you
> >>>
> >>> don''t
> >>>
> >>> >>> have any kernel for the guest system. In my case
I could only start
> >>>
> >>> a
> >>>
> >>> >>> xen
> >>> >>> guest (on FC7) with an older kernel-xenU
installed from FC version
> >>>
> >>> 5.
> >>>
> >>> >>> My question is: Why does the newer releases of
linux has xen kernel
> >>> >>> prebuilt but just for dom0, not for the guest
systems, and you
> >>> >>> can''t even
> >>> >>> find a domU kernel special for those systems?
> >>> >>>
> >>> >>> 2.My network topology in my home is like this:
> >>> >>> --------
> >>> >>> -router-
> >>> >>> --------
> >>> >>>
> >>> >>>
> >>> >>>
> >>> >>> -----------         ------------
> >>> >>> -linux xen-   ----  -other 2 pc-
> >>> >>> -----------         ------------
> >>> >>>
> >>> >>> The linux xen machine has two network interfaces
and xen installed.
> >>> >>> I want: -  one windows machine virtualized
> >>> >>>         -  one linux machine for which I want to
have a public
> >>> >>> ipaddress (to put the ip in DMS on the router)
and I want it to use
> >>> >>> eth0 (so in this case the traffic can not be
sniffed by other guest
> >>> >>> systems or dom0).
> >>> >>>
> >>> >>> ifconfig -a (on dom0) it shows like this:
> >>> >>> eth0      Link encap:Ethernet  HWaddr
00:00:E8:76:E2:4D
> >>> >>>           UP BROADCAST MULTICAST  MTU:1500 
Metric:1
> >>> >>>           RX packets:0 errors:0 dropped:0
overruns:0 frame:0
> >>> >>>           TX packets:0 errors:0 dropped:0
overruns:0 carrier:0
> >>> >>>           collisions:0 txqueuelen:1000
> >>> >>>           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
> >>> >>>           Interrupt:21 Base address:0x2000
> >>> >>>
> >>> >>> eth1      Link encap:Ethernet  HWaddr
00:16:76:B3:16:AB
> >>> >>>           inet addr:192.168.0.101 
Bcast:192.168.0.255
> >>> >>> Mask:255.255.255.0
> >>> >>>           inet6 addr: fe80::216:76ff:feb3:16ab/64
Scope:Link
> >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> >>> >>>           RX packets:198578 errors:0 dropped:0
overruns:0 frame:0
> >>> >>>           TX packets:117290 errors:0 dropped:0
overruns:0 carrier:0
> >>> >>>           collisions:0 txqueuelen:0
> >>> >>>           RX bytes:267328989 (254.9 MiB)  TX
bytes:8294632 (7.9
> >>> >>> MiB)
> >>> >>>
> >>> >>> lo        Link encap:Local Loopback
> >>> >>>           inet addr:127.0.0.1  Mask:255.0.0.0
> >>> >>>           inet6 addr: ::1/128 Scope:Host
> >>> >>>           UP LOOPBACK RUNNING  MTU:16436 
Metric:1
> >>> >>>           RX packets:2689 errors:0 dropped:0
overruns:0 frame:0
> >>> >>>           TX packets:2689 errors:0 dropped:0
overruns:0 carrier:0
> >>> >>>           collisions:0 txqueuelen:0
> >>> >>>           RX bytes:12510296 (11.9 MiB)  TX
bytes:12510296 (11.9
> >>> >>> MiB)
> >>> >>>
> >>> >>> peth1     Link encap:Ethernet  HWaddr
00:16:76:B3:16:AB
> >>> >>>           inet6 addr: fe80::216:76ff:feb3:16ab/64
Scope:Link
> >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> >>> >>>           RX packets:198588 errors:0 dropped:0
overruns:0 frame:0
> >>> >>>           TX packets:117311 errors:0 dropped:0
overruns:0 carrier:0
> >>> >>>           collisions:0 txqueuelen:100
> >>> >>>           RX bytes:270906777 (258.3 MiB)  TX
bytes:8813848 (8.4
> >>> >>> MiB) Base address:0x40c0 Memory:92200000-92220000
> >>> >>>
> >>> >>> vif4.0    Link encap:Ethernet  HWaddr
FE:FF:FF:FF:FF:FF
> >>> >>>           inet6 addr:
fe80::fcff:ffff:feff:ffff/64 Scope:Link
> >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> >>> >>>           RX packets:9 errors:0 dropped:0
overruns:0 frame:0
> >>> >>>           TX packets:1 errors:0 dropped:6
overruns:0 carrier:0
> >>> >>>           collisions:0 txqueuelen:1
> >>> >>>           RX bytes:1068 (1.0 KiB)  TX bytes:342
(342.0 b)
> >>> >>>
> >>> >>> virbr0    Link encap:Ethernet  HWaddr
FE:FF:FF:FF:FF:FF
> >>> >>>           inet addr:192.168.122.1 
Bcast:192.168.122.255
> >>> >>> Mask:255.255.255.0 inet6 addr:
fe80::200:ff:fe00:0/64 Scope:Link
> >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> >>> >>>           RX packets:43 errors:0 dropped:0
overruns:0 frame:0
> >>> >>>           TX packets:17 errors:0 dropped:0
overruns:0 carrier:0
> >>> >>>           collisions:0 txqueuelen:0
> >>> >>>           RX bytes:3208 (3.1 KiB)  TX bytes:2018
(1.9 KiB)
> >>> >>>
> >>> >>> I don''t see any xen bridge, because
that''s what I think I need: one
> >>> >>> network card, and one xen bridge.
> >>> >>> I found on google that I could use the following
script:
> >>> >>> #!/bin/sh
> >>> >>> dir=$(dirname "$0")
> >>> >>> "$dir/network-bridge" "$@"
vifnum=0 netdev=eth0 bridge=xenbr0
> >>> >>> "$dir/network-bridge" "$@"
vifnum=1 netdev=eth1 bridge=xenbr1
> >>> >>> "$dir/network-bridge" "$@"
vifnum=2 netdev=eth2 bridge=xenbr2
> >>> >>> And then set it into xen-config.sxp:
> >>> >>> network-script matrix-network
> >>> >>> But it gives errors that network-script has only
start, stop and
> >>> >>> status. The only thing that it succeds is that I
can see a xen
> >>>
> >>> bridge.
> >>>
> >>> >>> If this would work, doesn''t this affects
other guest domains also?
> >>> >>>
> >>> >>> My question is: How can I set a guest dom to use
directly a network
> >>> >>> card with other ip class ?
> >>> >>>
> >>> >>> Best regards,
> >>> >>> Octav
> >>> >>>
> >>> >>>
> >>> >>>
> >>> >>> _______________________________________________
> >>> >>> Xen-users mailing list
> >>> >>> Xen-users@lists.xensource.com
> >>> >>> http://lists.xensource.com/xen-users
> >>> >>
> >>> >> --
> >>> >> Kobalt W.I.T.
> >>> >> Web & Information Technology
> >>> >> Brusselsesteenweg 152
> >>> >> 1850 Grimbergen
> >>> >>
> >>> >> Tel  : +32 479 339 655
> >>> >> Email: info@kobaltwit.be
> >>> >>
> >>> >> _______________________________________________
> >>> >> Xen-users mailing list
> >>> >> Xen-users@lists.xensource.com
> >>> >> http://lists.xensource.com/xen-users
> >>
> >> --
> >> Kobalt W.I.T.
> >> Web & Information Technology
> >> Brusselsesteenweg 152
> >> 1850 Grimbergen
> >>
> >> Tel  : +32 479 339 655
> >> Email: info@kobaltwit.be
> >>
> >> _______________________________________________
> >> Xen-users mailing list
> >> Xen-users@lists.xensource.com
> >> http://lists.xensource.com/xen-users
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@lists.xensource.com
> http://lists.xensource.com/xen-users
-- 
Kobalt W.I.T.
Web & Information Technology
Brusselsesteenweg 152
1850 Grimbergen

Tel  : +32 479 339 655
Email: info@kobaltwit.be

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

Petersson, Mats

2007-Jun-05 12:28 UTC

head link

RE: [Xen-users] Xen a couple of questions

> -----Original Message-----
> From: xen-users-bounces@lists.xensource.com 
> [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of 
> Geert Janssens
> Sent: 05 June 2007 13:26
> To: xen-users@lists.xensource.com
> Cc: Octavian Teodorescu
> Subject: Re: [Xen-users] Xen a couple of questions
> 
> I am sorry, this is way over my head. I''m just a (relatively 
> new) user of xen 
> myself. I hope someone else can help you here.
> 
> Regards,
> 
> Geert
> 
> On Tuesday 5 June 2007 14:19, Octavian Teodorescu wrote:
> > I had stoped the virtual domain and modified in the config 
> file to use the
> > dom0 kernel.
> > Here are the last lines of the guest boot where you can see 
> the error:
> > "SCSI subsystem initialized
> > device-mapper: ioctl: 4.11.0-ioctl (2006-10-12) initialised:
> > dm-devel@redhat.com
> > Kernel panic - not syncing: Attempted to kill init!"
> >
> > But if I use the old kernelU i have which is actually for 
> Fedorac core 5
> > then everthing it''s ok.
Do you have an "ramdisk" entry in your PV domain config? You probably
need one. 

--
Mats> >
> > >> On Tuesday 5 June 2007 13:11, you wrote:
> > >>> Thanks a lot, I appreciate your help.
> > >>> 1. I tried with the same configuration and with the 
> kernel of dom0, but
> > >>> I
> > >>> receive a lot of errors, on both Fedora Core 7 and 
> Centos 5 (systems
> > >>> with
> > >>> which I''ve tried xen).
> > >>
> > >> Hmm, I don''t know what exactly you tried and what is
failing.
> > >>
> > >> I simply used virt-manager to create a CentOS 5 guest on 
> my CentOS 5
> > >> dom0. I
> > >> followed (more or less) the guidelines that come with CentOS
5''s
> > >> release:
http://www.centos.org/docs/5/html/Virtualization-en-US/
> > >>
> > >> I don''t remember having particular difficulty with
this.
> The only caveat
> > >> I remember was that Anaconda insists on a block device 
> that can hold a
> > >> partition map. So while installing, you can''t
provide
> the guest with
> > >> separate
> > >> partitions, because it will treat these separate 
> partitions as complete
> > >> diskss that have to be partitioned still.
> > >>
> > >> Regards,
> > >>
> > >> Geert
> > >>
> > >>> 2.Yap that''s a bridge interface. I''ll
look into your mails about
> > >>> advance bridging, thanks.
> > >>>
> > >>> Best regards.
> > >>>
> > >>> >> 1. On CentOS 5, Redhat Enterprise 5 and Fedora
core
> 6 and up, the
> > >>> >> xen kernel
> > >>> >> can be used for both dom0 and domU. There is no
need
> anymore for two
> > >>> >> kernels.
> > >>> >>
> > >>> >> 2. I don''t know the complete answer to
your second
> question. From
> > >>>
> > >>> your
> > >>>
> > >>> >> ifconfig output, it looks as if Fedora Core 7 
> creates a virbr0
> > >>>
> > >>> interface
> > >>>
> > >>> >> instead of a xenbr0 interface. You could check
if
> this is really a
> > >>> >> bridge with the command "brctl show".
You probably
> have to execute
> > >>>
> > >>> this
> > >>>
> > >>> >> as root.
> > >>> >>
> > >>> >> Then if I understand your question correctly,
you
> are trying to
> > >>> >> setup
> > >>>
> > >>> a
> > >>>
> > >>> >> xen
> > >>> >> guest domain to act as a 
> firewall/router/gateway/whatever for your
> > >>>
> > >>> lan.
> > >>>
> > >>> >> So I assume you only want this guest domain to
use
> the external
> > >>>
> > >>> network
> > >>>
> > >>> >> card
> > >>> >> (your eth0). There are two ways of accomplishing
this:
> > >>> >> * either use PCI passthrough so that your dom0
won''t
> see eth0, but
> > >>> >> instead it''s passed to your guest
system (search for
> pciback on
> > >>>
> > >>> Google
> > >>>
> > >>> >> for more info). Unfortunatly, I didn''t
manage to set
> this up in my
> > >>> >> particular case, so
> > >>> >> I used the seconde option:
> > >>> >> * create two xenbridges, one for your external 
> network interface,
> > >>> >> and one for
> > >>> >> your internal network interface. Then configure
dom0
> such that it
> > >>>
> > >>> isn''t
> > >>>
> > >>> >> allowed to use the bridge for the external 
> interface. You can do
> > >>> >> this
> > >>>
> > >>> by
> > >>>
> > >>> >> either disabling the virtual interface in dom0 
> (which will be called
> > >>> >> eth0) or
> > >>> >> by setting some firewall rules in dom0, or both.
> > >>> >> You can search this list for one of my earlier
mails, where I
> > >>> >> explain
> > >>>
> > >>> my
> > >>>
> > >>> >> configuration (on CentOS 5). It''s
titled "advanced
> bridging..." and
> > >>> >> dated May
> > >>> >> 16th, 2007.
> > >>> >>
> > >>> >> Hopefully this will help you along the way.
> > >>> >>
> > >>> >> Cheers,
> > >>> >>
> > >>> >> Geert
> > >>> >>
> > >>> >> On Tuesday 5 June 2007 10:34, Octavian
Teodorescu wrote:
> > >>> >>> Hi guys,
> > >>> >>>
> > >>> >>> 1. Regarding Centos and Fedora core 7
compared with
> fedora core 5.
> > >>>
> > >>> I''ve
> > >>>
> > >>> >>> seen that on fedora core 5 when you want to
install
> xen you have to
> > >>> >>> install the following packages: xen,
kernel-xen0
> and kernel-xenU
> > >>> >>> (of course with the dependencies needed).
But on
> Centos, FC7 and I
> > >>> >>> think redhat versions, you only have to
install xen
> and kernel-xen,
> > >>> >>> you
> > >>>
> > >>> don''t
> > >>>
> > >>> >>> have any kernel for the guest system. In my
case I
> could only start
> > >>>
> > >>> a
> > >>>
> > >>> >>> xen
> > >>> >>> guest (on FC7) with an older kernel-xenU
installed
> from FC version
> > >>>
> > >>> 5.
> > >>>
> > >>> >>> My question is: Why does the newer releases
of
> linux has xen kernel
> > >>> >>> prebuilt but just for dom0, not for the
guest
> systems, and you
> > >>> >>> can''t even
> > >>> >>> find a domU kernel special for those
systems?
> > >>> >>>
> > >>> >>> 2.My network topology in my home is like
this:
> > >>> >>> --------
> > >>> >>> -router-
> > >>> >>> --------
> > >>> >>>
> > >>> >>>
> > >>> >>>
> > >>> >>> -----------         ------------
> > >>> >>> -linux xen-   ----  -other 2 pc-
> > >>> >>> -----------         ------------
> > >>> >>>
> > >>> >>> The linux xen machine has two network
interfaces
> and xen installed.
> > >>> >>> I want: -  one windows machine virtualized
> > >>> >>>         -  one linux machine for which I
want to
> have a public
> > >>> >>> ipaddress (to put the ip in DMS on the
router) and
> I want it to use
> > >>> >>> eth0 (so in this case the traffic can not be
> sniffed by other guest
> > >>> >>> systems or dom0).
> > >>> >>>
> > >>> >>> ifconfig -a (on dom0) it shows like this:
> > >>> >>> eth0      Link encap:Ethernet  HWaddr
00:00:E8:76:E2:4D
> > >>> >>>           UP BROADCAST MULTICAST  MTU:1500 
Metric:1
> > >>> >>>           RX packets:0 errors:0 dropped:0
overruns:0 frame:0
> > >>> >>>           TX packets:0 errors:0 dropped:0 
> overruns:0 carrier:0
> > >>> >>>           collisions:0 txqueuelen:1000
> > >>> >>>           RX bytes:0 (0.0 b)  TX bytes:0
(0.0 b)
> > >>> >>>           Interrupt:21 Base address:0x2000
> > >>> >>>
> > >>> >>> eth1      Link encap:Ethernet  HWaddr
00:16:76:B3:16:AB
> > >>> >>>           inet addr:192.168.0.101 
Bcast:192.168.0.255
> > >>> >>> Mask:255.255.255.0
> > >>> >>>           inet6 addr:
fe80::216:76ff:feb3:16ab/64 Scope:Link
> > >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> > >>> >>>           RX packets:198578 errors:0
dropped:0
> overruns:0 frame:0
> > >>> >>>           TX packets:117290 errors:0
dropped:0
> overruns:0 carrier:0
> > >>> >>>           collisions:0 txqueuelen:0
> > >>> >>>           RX bytes:267328989 (254.9 MiB)  TX
> bytes:8294632 (7.9
> > >>> >>> MiB)
> > >>> >>>
> > >>> >>> lo        Link encap:Local Loopback
> > >>> >>>           inet addr:127.0.0.1 
Mask:255.0.0.0
> > >>> >>>           inet6 addr: ::1/128 Scope:Host
> > >>> >>>           UP LOOPBACK RUNNING  MTU:16436 
Metric:1
> > >>> >>>           RX packets:2689 errors:0 dropped:0
> overruns:0 frame:0
> > >>> >>>           TX packets:2689 errors:0 dropped:0
> overruns:0 carrier:0
> > >>> >>>           collisions:0 txqueuelen:0
> > >>> >>>           RX bytes:12510296 (11.9 MiB)  TX 
> bytes:12510296 (11.9
> > >>> >>> MiB)
> > >>> >>>
> > >>> >>> peth1     Link encap:Ethernet  HWaddr
00:16:76:B3:16:AB
> > >>> >>>           inet6 addr:
fe80::216:76ff:feb3:16ab/64 Scope:Link
> > >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> > >>> >>>           RX packets:198588 errors:0
dropped:0
> overruns:0 frame:0
> > >>> >>>           TX packets:117311 errors:0
dropped:0
> overruns:0 carrier:0
> > >>> >>>           collisions:0 txqueuelen:100
> > >>> >>>           RX bytes:270906777 (258.3 MiB)  TX
> bytes:8813848 (8.4
> > >>> >>> MiB) Base address:0x40c0
Memory:92200000-92220000
> > >>> >>>
> > >>> >>> vif4.0    Link encap:Ethernet  HWaddr
FE:FF:FF:FF:FF:FF
> > >>> >>>           inet6 addr:
fe80::fcff:ffff:feff:ffff/64
> Scope:Link
> > >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> > >>> >>>           RX packets:9 errors:0 dropped:0
overruns:0 frame:0
> > >>> >>>           TX packets:1 errors:0 dropped:6 
> overruns:0 carrier:0
> > >>> >>>           collisions:0 txqueuelen:1
> > >>> >>>           RX bytes:1068 (1.0 KiB)  TX
bytes:342 (342.0 b)
> > >>> >>>
> > >>> >>> virbr0    Link encap:Ethernet  HWaddr
FE:FF:FF:FF:FF:FF
> > >>> >>>           inet addr:192.168.122.1 
Bcast:192.168.122.255
> > >>> >>> Mask:255.255.255.0 inet6 addr: 
> fe80::200:ff:fe00:0/64 Scope:Link
> > >>> >>>           UP BROADCAST RUNNING MULTICAST 
MTU:1500  Metric:1
> > >>> >>>           RX packets:43 errors:0 dropped:0 
> overruns:0 frame:0
> > >>> >>>           TX packets:17 errors:0 dropped:0 
> overruns:0 carrier:0
> > >>> >>>           collisions:0 txqueuelen:0
> > >>> >>>           RX bytes:3208 (3.1 KiB)  TX
bytes:2018 (1.9 KiB)
> > >>> >>>
> > >>> >>> I don''t see any xen bridge, because
that''s what I
> think I need: one
> > >>> >>> network card, and one xen bridge.
> > >>> >>> I found on google that I could use the
following script:
> > >>> >>> #!/bin/sh
> > >>> >>> dir=$(dirname "$0")
> > >>> >>> "$dir/network-bridge"
"$@" vifnum=0 netdev=eth0
> bridge=xenbr0
> > >>> >>> "$dir/network-bridge"
"$@" vifnum=1 netdev=eth1
> bridge=xenbr1
> > >>> >>> "$dir/network-bridge"
"$@" vifnum=2 netdev=eth2
> bridge=xenbr2
> > >>> >>> And then set it into xen-config.sxp:
> > >>> >>> network-script matrix-network
> > >>> >>> But it gives errors that network-script has
only
> start, stop and
> > >>> >>> status. The only thing that it succeds is
that I
> can see a xen
> > >>>
> > >>> bridge.
> > >>>
> > >>> >>> If this would work, doesn''t this
affects other
> guest domains also?
> > >>> >>>
> > >>> >>> My question is: How can I set a guest dom to
use
> directly a network
> > >>> >>> card with other ip class ?
> > >>> >>>
> > >>> >>> Best regards,
> > >>> >>> Octav
> > >>> >>>
> > >>> >>>
> > >>> >>>
> > >>> >>>
_______________________________________________
> > >>> >>> Xen-users mailing list
> > >>> >>> Xen-users@lists.xensource.com
> > >>> >>> http://lists.xensource.com/xen-users
> > >>> >>
> > >>> >> --
> > >>> >> Kobalt W.I.T.
> > >>> >> Web & Information Technology
> > >>> >> Brusselsesteenweg 152
> > >>> >> 1850 Grimbergen
> > >>> >>
> > >>> >> Tel  : +32 479 339 655
> > >>> >> Email: info@kobaltwit.be
> > >>> >>
> > >>> >> _______________________________________________
> > >>> >> Xen-users mailing list
> > >>> >> Xen-users@lists.xensource.com
> > >>> >> http://lists.xensource.com/xen-users
> > >>
> > >> --
> > >> Kobalt W.I.T.
> > >> Web & Information Technology
> > >> Brusselsesteenweg 152
> > >> 1850 Grimbergen
> > >>
> > >> Tel  : +32 479 339 655
> > >> Email: info@kobaltwit.be
> > >>
> > >> _______________________________________________
> > >> Xen-users mailing list
> > >> Xen-users@lists.xensource.com
> > >> http://lists.xensource.com/xen-users
> >
> > _______________________________________________
> > Xen-users mailing list
> > Xen-users@lists.xensource.com
> > http://lists.xensource.com/xen-users
> 
> -- 
> Kobalt W.I.T.
> Web & Information Technology
> Brusselsesteenweg 152
> 1850 Grimbergen
> 
> Tel  : +32 479 339 655
> Email: info@kobaltwit.be
> 
> _______________________________________________
> Xen-users mailing list
> Xen-users@lists.xensource.com
> http://lists.xensource.com/xen-users
> 
> 
> 


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

Octavian Teodorescu

2007-Jun-05 12:50 UTC

head link

RE: [Xen-users] Xen a couple of questions

Yap, I have the ramdisk of dom0 also. Without the ramdisk set up, I''ve
seen that it can''t mount logical volumes as partitions (with any kind
of
kernel).

Best regards.
>>
>>
>>> -----Original Message-----
>>> From: xen-users-bounces@lists.xensource.com
>>> [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of
>>> Geert Janssens
>>> Sent: 05 June 2007 13:26
>>> To: xen-users@lists.xensource.com
>>> Cc: Octavian Teodorescu
>>> Subject: Re: [Xen-users] Xen a couple of questions
>>>
>>> I am sorry, this is way over my head. I''m just a
(relatively
>>> new) user of xen
>>> myself. I hope someone else can help you here.
>>>
>>> Regards,
>>>
>>> Geert
>>>
>>> On Tuesday 5 June 2007 14:19, Octavian Teodorescu wrote:
>>> > I had stoped the virtual domain and modified in the config
>>> file to use the
>>> > dom0 kernel.
>>> > Here are the last lines of the guest boot where you can see
>>> the error:
>>> > "SCSI subsystem initialized
>>> > device-mapper: ioctl: 4.11.0-ioctl (2006-10-12) initialised:
>>> > dm-devel@redhat.com
>>> > Kernel panic - not syncing: Attempted to kill init!"
>>> >
>>> > But if I use the old kernelU i have which is actually for
>>> Fedorac core 5
>>> > then everthing it''s ok.
>>
>> Do you have an "ramdisk" entry in your PV domain config? You
probably
>> need one.
>>
>> --
>> Mats
>>> >
>>> > >> On Tuesday 5 June 2007 13:11, you wrote:
>>> > >>> Thanks a lot, I appreciate your help.
>>> > >>> 1. I tried with the same configuration and with
the
>>> kernel of dom0, but
>>> > >>> I
>>> > >>> receive a lot of errors, on both Fedora Core 7
and
>>> Centos 5 (systems
>>> > >>> with
>>> > >>> which I''ve tried xen).
>>> > >>
>>> > >> Hmm, I don''t know what exactly you tried and
what is failing.
>>> > >>
>>> > >> I simply used virt-manager to create a CentOS 5 guest
on
>>> my CentOS 5
>>> > >> dom0. I
>>> > >> followed (more or less) the guidelines that come with
CentOS 5''s
>>> > >> release:
http://www.centos.org/docs/5/html/Virtualization-en-US/
>>> > >>
>>> > >> I don''t remember having particular
difficulty with this.
>>> The only caveat
>>> > >> I remember was that Anaconda insists on a block
device
>>> that can hold a
>>> > >> partition map. So while installing, you
can''t provide
>>> the guest with
>>> > >> separate
>>> > >> partitions, because it will treat these separate
>>> partitions as complete
>>> > >> diskss that have to be partitioned still.
>>> > >>
>>> > >> Regards,
>>> > >>
>>> > >> Geert
>>> > >>
>>> > >>> 2.Yap that''s a bridge interface.
I''ll look into your mails about
>>> > >>> advance bridging, thanks.
>>> > >>>
>>> > >>> Best regards.
>>> > >>>
>>> > >>> >> 1. On CentOS 5, Redhat Enterprise 5 and
Fedora core
>>> 6 and up, the
>>> > >>> >> xen kernel
>>> > >>> >> can be used for both dom0 and domU.
There is no need
>>> anymore for two
>>> > >>> >> kernels.
>>> > >>> >>
>>> > >>> >> 2. I don''t know the complete
answer to your second
>>> question. From
>>> > >>>
>>> > >>> your
>>> > >>>
>>> > >>> >> ifconfig output, it looks as if Fedora
Core 7
>>> creates a virbr0
>>> > >>>
>>> > >>> interface
>>> > >>>
>>> > >>> >> instead of a xenbr0 interface. You could
check if
>>> this is really a
>>> > >>> >> bridge with the command "brctl
show". You probably
>>> have to execute
>>> > >>>
>>> > >>> this
>>> > >>>
>>> > >>> >> as root.
>>> > >>> >>
>>> > >>> >> Then if I understand your question
correctly, you
>>> are trying to
>>> > >>> >> setup
>>> > >>>
>>> > >>> a
>>> > >>>
>>> > >>> >> xen
>>> > >>> >> guest domain to act as a
>>> firewall/router/gateway/whatever for your
>>> > >>>
>>> > >>> lan.
>>> > >>>
>>> > >>> >> So I assume you only want this guest
domain to use
>>> the external
>>> > >>>
>>> > >>> network
>>> > >>>
>>> > >>> >> card
>>> > >>> >> (your eth0). There are two ways of
accomplishing this:
>>> > >>> >> * either use PCI passthrough so that
your dom0 won''t
>>> see eth0, but
>>> > >>> >> instead it''s passed to your
guest system (search for
>>> pciback on
>>> > >>>
>>> > >>> Google
>>> > >>>
>>> > >>> >> for more info). Unfortunatly, I
didn''t manage to set
>>> this up in my
>>> > >>> >> particular case, so
>>> > >>> >> I used the seconde option:
>>> > >>> >> * create two xenbridges, one for your
external
>>> network interface,
>>> > >>> >> and one for
>>> > >>> >> your internal network interface. Then
configure dom0
>>> such that it
>>> > >>>
>>> > >>> isn''t
>>> > >>>
>>> > >>> >> allowed to use the bridge for the
external
>>> interface. You can do
>>> > >>> >> this
>>> > >>>
>>> > >>> by
>>> > >>>
>>> > >>> >> either disabling the virtual interface
in dom0
>>> (which will be called
>>> > >>> >> eth0) or
>>> > >>> >> by setting some firewall rules in dom0,
or both.
>>> > >>> >> You can search this list for one of my
earlier mails, where I
>>> > >>> >> explain
>>> > >>>
>>> > >>> my
>>> > >>>
>>> > >>> >> configuration (on CentOS 5).
It''s titled "advanced
>>> bridging..." and
>>> > >>> >> dated May
>>> > >>> >> 16th, 2007.
>>> > >>> >>
>>> > >>> >> Hopefully this will help you along the
way.
>>> > >>> >>
>>> > >>> >> Cheers,
>>> > >>> >>
>>> > >>> >> Geert
>>> > >>> >>
>>> > >>> >> On Tuesday 5 June 2007 10:34, Octavian
Teodorescu wrote:
>>> > >>> >>> Hi guys,
>>> > >>> >>>
>>> > >>> >>> 1. Regarding Centos and Fedora core
7 compared with
>>> fedora core 5.
>>> > >>>
>>> > >>> I''ve
>>> > >>>
>>> > >>> >>> seen that on fedora core 5 when you
want to install
>>> xen you have to
>>> > >>> >>> install the following packages: xen,
kernel-xen0
>>> and kernel-xenU
>>> > >>> >>> (of course with the dependencies
needed). But on
>>> Centos, FC7 and I
>>> > >>> >>> think redhat versions, you only have
to install xen
>>> and kernel-xen,
>>> > >>> >>> you
>>> > >>>
>>> > >>> don''t
>>> > >>>
>>> > >>> >>> have any kernel for the guest
system. In my case I
>>> could only start
>>> > >>>
>>> > >>> a
>>> > >>>
>>> > >>> >>> xen
>>> > >>> >>> guest (on FC7) with an older
kernel-xenU installed
>>> from FC version
>>> > >>>
>>> > >>> 5.
>>> > >>>
>>> > >>> >>> My question is: Why does the newer
releases of
>>> linux has xen kernel
>>> > >>> >>> prebuilt but just for dom0, not for
the guest
>>> systems, and you
>>> > >>> >>> can''t even
>>> > >>> >>> find a domU kernel special for those
systems?
>>> > >>> >>>
>>> > >>> >>> 2.My network topology in my home is
like this:
>>> > >>> >>> --------
>>> > >>> >>> -router-
>>> > >>> >>> --------
>>> > >>> >>>
>>> > >>> >>>
>>> > >>> >>>
>>> > >>> >>> -----------         ------------
>>> > >>> >>> -linux xen-   ----  -other 2 pc-
>>> > >>> >>> -----------         ------------
>>> > >>> >>>
>>> > >>> >>> The linux xen machine has two
network interfaces
>>> and xen installed.
>>> > >>> >>> I want: -  one windows machine
virtualized
>>> > >>> >>>         -  one linux machine for
which I want to
>>> have a public
>>> > >>> >>> ipaddress (to put the ip in DMS on
the router) and
>>> I want it to use
>>> > >>> >>> eth0 (so in this case the traffic
can not be
>>> sniffed by other guest
>>> > >>> >>> systems or dom0).
>>> > >>> >>>
>>> > >>> >>> ifconfig -a (on dom0) it shows like
this:
>>> > >>> >>> eth0      Link encap:Ethernet 
HWaddr 00:00:E8:76:E2:4D
>>> > >>> >>>           UP BROADCAST MULTICAST 
MTU:1500  Metric:1
>>> > >>> >>>           RX packets:0 errors:0
dropped:0 overruns:0 frame:0
>>> > >>> >>>           TX packets:0 errors:0
dropped:0
>>> overruns:0 carrier:0
>>> > >>> >>>           collisions:0
txqueuelen:1000
>>> > >>> >>>           RX bytes:0 (0.0 b)  TX
bytes:0 (0.0 b)
>>> > >>> >>>           Interrupt:21 Base
address:0x2000
>>> > >>> >>>
>>> > >>> >>> eth1      Link encap:Ethernet 
HWaddr 00:16:76:B3:16:AB
>>> > >>> >>>           inet addr:192.168.0.101 
Bcast:192.168.0.255
>>> > >>> >>> Mask:255.255.255.0
>>> > >>> >>>           inet6 addr:
fe80::216:76ff:feb3:16ab/64 Scope:Link
>>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST  MTU:1500  Metric:1
>>> > >>> >>>           RX packets:198578 errors:0
dropped:0
>>> overruns:0 frame:0
>>> > >>> >>>           TX packets:117290 errors:0
dropped:0
>>> overruns:0 carrier:0
>>> > >>> >>>           collisions:0 txqueuelen:0
>>> > >>> >>>           RX bytes:267328989 (254.9
MiB)  TX
>>> bytes:8294632 (7.9
>>> > >>> >>> MiB)
>>> > >>> >>>
>>> > >>> >>> lo        Link encap:Local Loopback
>>> > >>> >>>           inet addr:127.0.0.1 
Mask:255.0.0.0
>>> > >>> >>>           inet6 addr: ::1/128
Scope:Host
>>> > >>> >>>           UP LOOPBACK RUNNING 
MTU:16436  Metric:1
>>> > >>> >>>           RX packets:2689 errors:0
dropped:0
>>> overruns:0 frame:0
>>> > >>> >>>           TX packets:2689 errors:0
dropped:0
>>> overruns:0 carrier:0
>>> > >>> >>>           collisions:0 txqueuelen:0
>>> > >>> >>>           RX bytes:12510296 (11.9
MiB)  TX
>>> bytes:12510296 (11.9
>>> > >>> >>> MiB)
>>> > >>> >>>
>>> > >>> >>> peth1     Link encap:Ethernet 
HWaddr 00:16:76:B3:16:AB
>>> > >>> >>>           inet6 addr:
fe80::216:76ff:feb3:16ab/64 Scope:Link
>>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST  MTU:1500  Metric:1
>>> > >>> >>>           RX packets:198588 errors:0
dropped:0
>>> overruns:0 frame:0
>>> > >>> >>>           TX packets:117311 errors:0
dropped:0
>>> overruns:0 carrier:0
>>> > >>> >>>           collisions:0
txqueuelen:100
>>> > >>> >>>           RX bytes:270906777 (258.3
MiB)  TX
>>> bytes:8813848 (8.4
>>> > >>> >>> MiB) Base address:0x40c0
Memory:92200000-92220000
>>> > >>> >>>
>>> > >>> >>> vif4.0    Link encap:Ethernet 
HWaddr FE:FF:FF:FF:FF:FF
>>> > >>> >>>           inet6 addr:
fe80::fcff:ffff:feff:ffff/64
>>> Scope:Link
>>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST  MTU:1500  Metric:1
>>> > >>> >>>           RX packets:9 errors:0
dropped:0 overruns:0 frame:0
>>> > >>> >>>           TX packets:1 errors:0
dropped:6
>>> overruns:0 carrier:0
>>> > >>> >>>           collisions:0 txqueuelen:1
>>> > >>> >>>           RX bytes:1068 (1.0 KiB) 
TX bytes:342 (342.0 b)
>>> > >>> >>>
>>> > >>> >>> virbr0    Link encap:Ethernet 
HWaddr FE:FF:FF:FF:FF:FF
>>> > >>> >>>           inet addr:192.168.122.1 
Bcast:192.168.122.255
>>> > >>> >>> Mask:255.255.255.0 inet6 addr:
>>> fe80::200:ff:fe00:0/64 Scope:Link
>>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST  MTU:1500  Metric:1
>>> > >>> >>>           RX packets:43 errors:0
dropped:0
>>> overruns:0 frame:0
>>> > >>> >>>           TX packets:17 errors:0
dropped:0
>>> overruns:0 carrier:0
>>> > >>> >>>           collisions:0 txqueuelen:0
>>> > >>> >>>           RX bytes:3208 (3.1 KiB) 
TX bytes:2018 (1.9 KiB)
>>> > >>> >>>
>>> > >>> >>> I don''t see any xen bridge,
because that''s what I
>>> think I need: one
>>> > >>> >>> network card, and one xen bridge.
>>> > >>> >>> I found on google that I could use
the following script:
>>> > >>> >>> #!/bin/sh
>>> > >>> >>> dir=$(dirname "$0")
>>> > >>> >>> "$dir/network-bridge"
"$@" vifnum=0 netdev=eth0
>>> bridge=xenbr0
>>> > >>> >>> "$dir/network-bridge"
"$@" vifnum=1 netdev=eth1
>>> bridge=xenbr1
>>> > >>> >>> "$dir/network-bridge"
"$@" vifnum=2 netdev=eth2
>>> bridge=xenbr2
>>> > >>> >>> And then set it into xen-config.sxp:
>>> > >>> >>> network-script matrix-network
>>> > >>> >>> But it gives errors that
network-script has only
>>> start, stop and
>>> > >>> >>> status. The only thing that it
succeds is that I
>>> can see a xen
>>> > >>>
>>> > >>> bridge.
>>> > >>>
>>> > >>> >>> If this would work, doesn''t
this affects other
>>> guest domains also?
>>> > >>> >>>
>>> > >>> >>> My question is: How can I set a
guest dom to use
>>> directly a network
>>> > >>> >>> card with other ip class ?
>>> > >>> >>>
>>> > >>> >>> Best regards,
>>> > >>> >>> Octav
>>> > >>> >>>
>>> > >>> >>>
>>> > >>> >>>
>>> > >>> >>>
_______________________________________________
>>> > >>> >>> Xen-users mailing list
>>> > >>> >>> Xen-users@lists.xensource.com
>>> > >>> >>> http://lists.xensource.com/xen-users
>>> > >>> >>
>>> > >>> >> --
>>> > >>> >> Kobalt W.I.T.
>>> > >>> >> Web & Information Technology
>>> > >>> >> Brusselsesteenweg 152
>>> > >>> >> 1850 Grimbergen
>>> > >>> >>
>>> > >>> >> Tel  : +32 479 339 655
>>> > >>> >> Email: info@kobaltwit.be
>>> > >>> >>
>>> > >>> >>
_______________________________________________
>>> > >>> >> Xen-users mailing list
>>> > >>> >> Xen-users@lists.xensource.com
>>> > >>> >> http://lists.xensource.com/xen-users
>>> > >>
>>> > >> --
>>> > >> Kobalt W.I.T.
>>> > >> Web & Information Technology
>>> > >> Brusselsesteenweg 152
>>> > >> 1850 Grimbergen
>>> > >>
>>> > >> Tel  : +32 479 339 655
>>> > >> Email: info@kobaltwit.be
>>> > >>
>>> > >> _______________________________________________
>>> > >> Xen-users mailing list
>>> > >> Xen-users@lists.xensource.com
>>> > >> http://lists.xensource.com/xen-users
>>> >
>>> > _______________________________________________
>>> > Xen-users mailing list
>>> > Xen-users@lists.xensource.com
>>> > http://lists.xensource.com/xen-users
>>>
>>> --
>>> Kobalt W.I.T.
>>> Web & Information Technology
>>> Brusselsesteenweg 152
>>> 1850 Grimbergen
>>>
>>> Tel  : +32 479 339 655
>>> Email: info@kobaltwit.be
>>>
>>> _______________________________________________
>>> Xen-users mailing list
>>> Xen-users@lists.xensource.com
>>> http://lists.xensource.com/xen-users
>>>
>>>
>>>
>>
>>
>>
>> _______________________________________________
>> Xen-users mailing list
>> Xen-users@lists.xensource.com
>> http://lists.xensource.com/xen-users
>>


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

Petersson, Mats

2007-Jun-05 12:54 UTC

head link

RE: [Xen-users] Xen a couple of questions

> -----Original Message-----
> From: Octavian Teodorescu [mailto:octav@inetsolutions.ro] 
> Sent: 05 June 2007 13:50
> To: Petersson, Mats
> Cc: xen-users@lists.xensource.com
> Subject: RE: [Xen-users] Xen a couple of questions
> 
> Yap, I have the ramdisk of dom0 also. Without the ramdisk set up,
I''ve
> seen that it can''t mount logical volumes as partitions (with 
> any kind of
> kernel).
Yes, but you need one that uses blkfront driver if you are using it in
domU''s. You can technically create the same one for both DomU and Dom0,
but I think it''s easier/better to have separate ones, as at least you
don''t risk breaking the working Dom0 one trying to set the DomU one up.

--
Mats> 
> Best regards.
> 
> >>
> >>
> >>> -----Original Message-----
> >>> From: xen-users-bounces@lists.xensource.com
> >>> [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of
> >>> Geert Janssens
> >>> Sent: 05 June 2007 13:26
> >>> To: xen-users@lists.xensource.com
> >>> Cc: Octavian Teodorescu
> >>> Subject: Re: [Xen-users] Xen a couple of questions
> >>>
> >>> I am sorry, this is way over my head. I''m just a
(relatively
> >>> new) user of xen
> >>> myself. I hope someone else can help you here.
> >>>
> >>> Regards,
> >>>
> >>> Geert
> >>>
> >>> On Tuesday 5 June 2007 14:19, Octavian Teodorescu wrote:
> >>> > I had stoped the virtual domain and modified in the
config
> >>> file to use the
> >>> > dom0 kernel.
> >>> > Here are the last lines of the guest boot where you can
see
> >>> the error:
> >>> > "SCSI subsystem initialized
> >>> > device-mapper: ioctl: 4.11.0-ioctl (2006-10-12)
initialised:
> >>> > dm-devel@redhat.com
> >>> > Kernel panic - not syncing: Attempted to kill init!"
> >>> >
> >>> > But if I use the old kernelU i have which is actually for
> >>> Fedorac core 5
> >>> > then everthing it''s ok.
> >>
> >> Do you have an "ramdisk" entry in your PV domain config?
> You probably
> >> need one.
> >>
> >> --
> >> Mats
> >>> >
> >>> > >> On Tuesday 5 June 2007 13:11, you wrote:
> >>> > >>> Thanks a lot, I appreciate your help.
> >>> > >>> 1. I tried with the same configuration and
with the
> >>> kernel of dom0, but
> >>> > >>> I
> >>> > >>> receive a lot of errors, on both Fedora Core
7 and
> >>> Centos 5 (systems
> >>> > >>> with
> >>> > >>> which I''ve tried xen).
> >>> > >>
> >>> > >> Hmm, I don''t know what exactly you
tried and what is failing.
> >>> > >>
> >>> > >> I simply used virt-manager to create a CentOS 5
guest on
> >>> my CentOS 5
> >>> > >> dom0. I
> >>> > >> followed (more or less) the guidelines that come
> with CentOS 5''s
> >>> > >> release: 
> http://www.centos.org/docs/5/html/Virtualization-en-US/
> >>> > >>
> >>> > >> I don''t remember having particular
difficulty with this.
> >>> The only caveat
> >>> > >> I remember was that Anaconda insists on a block
device
> >>> that can hold a
> >>> > >> partition map. So while installing, you
can''t provide
> >>> the guest with
> >>> > >> separate
> >>> > >> partitions, because it will treat these separate
> >>> partitions as complete
> >>> > >> diskss that have to be partitioned still.
> >>> > >>
> >>> > >> Regards,
> >>> > >>
> >>> > >> Geert
> >>> > >>
> >>> > >>> 2.Yap that''s a bridge interface.
I''ll look into
> your mails about
> >>> > >>> advance bridging, thanks.
> >>> > >>>
> >>> > >>> Best regards.
> >>> > >>>
> >>> > >>> >> 1. On CentOS 5, Redhat Enterprise 5
and Fedora core
> >>> 6 and up, the
> >>> > >>> >> xen kernel
> >>> > >>> >> can be used for both dom0 and domU.
There is no need
> >>> anymore for two
> >>> > >>> >> kernels.
> >>> > >>> >>
> >>> > >>> >> 2. I don''t know the
complete answer to your second
> >>> question. From
> >>> > >>>
> >>> > >>> your
> >>> > >>>
> >>> > >>> >> ifconfig output, it looks as if
Fedora Core 7
> >>> creates a virbr0
> >>> > >>>
> >>> > >>> interface
> >>> > >>>
> >>> > >>> >> instead of a xenbr0 interface. You
could check if
> >>> this is really a
> >>> > >>> >> bridge with the command "brctl
show". You probably
> >>> have to execute
> >>> > >>>
> >>> > >>> this
> >>> > >>>
> >>> > >>> >> as root.
> >>> > >>> >>
> >>> > >>> >> Then if I understand your question
correctly, you
> >>> are trying to
> >>> > >>> >> setup
> >>> > >>>
> >>> > >>> a
> >>> > >>>
> >>> > >>> >> xen
> >>> > >>> >> guest domain to act as a
> >>> firewall/router/gateway/whatever for your
> >>> > >>>
> >>> > >>> lan.
> >>> > >>>
> >>> > >>> >> So I assume you only want this
guest domain to use
> >>> the external
> >>> > >>>
> >>> > >>> network
> >>> > >>>
> >>> > >>> >> card
> >>> > >>> >> (your eth0). There are two ways of
accomplishing this:
> >>> > >>> >> * either use PCI passthrough so
that your dom0 won''t
> >>> see eth0, but
> >>> > >>> >> instead it''s passed to
your guest system (search for
> >>> pciback on
> >>> > >>>
> >>> > >>> Google
> >>> > >>>
> >>> > >>> >> for more info). Unfortunatly, I
didn''t manage to set
> >>> this up in my
> >>> > >>> >> particular case, so
> >>> > >>> >> I used the seconde option:
> >>> > >>> >> * create two xenbridges, one for
your external
> >>> network interface,
> >>> > >>> >> and one for
> >>> > >>> >> your internal network interface.
Then configure dom0
> >>> such that it
> >>> > >>>
> >>> > >>> isn''t
> >>> > >>>
> >>> > >>> >> allowed to use the bridge for the
external
> >>> interface. You can do
> >>> > >>> >> this
> >>> > >>>
> >>> > >>> by
> >>> > >>>
> >>> > >>> >> either disabling the virtual
interface in dom0
> >>> (which will be called
> >>> > >>> >> eth0) or
> >>> > >>> >> by setting some firewall rules in
dom0, or both.
> >>> > >>> >> You can search this list for one of
my earlier
> mails, where I
> >>> > >>> >> explain
> >>> > >>>
> >>> > >>> my
> >>> > >>>
> >>> > >>> >> configuration (on CentOS 5).
It''s titled "advanced
> >>> bridging..." and
> >>> > >>> >> dated May
> >>> > >>> >> 16th, 2007.
> >>> > >>> >>
> >>> > >>> >> Hopefully this will help you along
the way.
> >>> > >>> >>
> >>> > >>> >> Cheers,
> >>> > >>> >>
> >>> > >>> >> Geert
> >>> > >>> >>
> >>> > >>> >> On Tuesday 5 June 2007 10:34,
Octavian Teodorescu wrote:
> >>> > >>> >>> Hi guys,
> >>> > >>> >>>
> >>> > >>> >>> 1. Regarding Centos and Fedora
core 7 compared with
> >>> fedora core 5.
> >>> > >>>
> >>> > >>> I''ve
> >>> > >>>
> >>> > >>> >>> seen that on fedora core 5 when
you want to install
> >>> xen you have to
> >>> > >>> >>> install the following packages:
xen, kernel-xen0
> >>> and kernel-xenU
> >>> > >>> >>> (of course with the
dependencies needed). But on
> >>> Centos, FC7 and I
> >>> > >>> >>> think redhat versions, you only
have to install xen
> >>> and kernel-xen,
> >>> > >>> >>> you
> >>> > >>>
> >>> > >>> don''t
> >>> > >>>
> >>> > >>> >>> have any kernel for the guest
system. In my case I
> >>> could only start
> >>> > >>>
> >>> > >>> a
> >>> > >>>
> >>> > >>> >>> xen
> >>> > >>> >>> guest (on FC7) with an older
kernel-xenU installed
> >>> from FC version
> >>> > >>>
> >>> > >>> 5.
> >>> > >>>
> >>> > >>> >>> My question is: Why does the
newer releases of
> >>> linux has xen kernel
> >>> > >>> >>> prebuilt but just for dom0, not
for the guest
> >>> systems, and you
> >>> > >>> >>> can''t even
> >>> > >>> >>> find a domU kernel special for
those systems?
> >>> > >>> >>>
> >>> > >>> >>> 2.My network topology in my
home is like this:
> >>> > >>> >>> --------
> >>> > >>> >>> -router-
> >>> > >>> >>> --------
> >>> > >>> >>>
> >>> > >>> >>>
> >>> > >>> >>>
> >>> > >>> >>> -----------        
------------
> >>> > >>> >>> -linux xen-   ----  -other 2
pc-
> >>> > >>> >>> -----------        
------------
> >>> > >>> >>>
> >>> > >>> >>> The linux xen machine has two
network interfaces
> >>> and xen installed.
> >>> > >>> >>> I want: -  one windows machine
virtualized
> >>> > >>> >>>         -  one linux machine
for which I want to
> >>> have a public
> >>> > >>> >>> ipaddress (to put the ip in DMS
on the router) and
> >>> I want it to use
> >>> > >>> >>> eth0 (so in this case the
traffic can not be
> >>> sniffed by other guest
> >>> > >>> >>> systems or dom0).
> >>> > >>> >>>
> >>> > >>> >>> ifconfig -a (on dom0) it shows
like this:
> >>> > >>> >>> eth0      Link encap:Ethernet 
HWaddr 00:00:E8:76:E2:4D
> >>> > >>> >>>           UP BROADCAST
MULTICAST  MTU:1500  Metric:1
> >>> > >>> >>>           RX packets:0 errors:0
dropped:0
> overruns:0 frame:0
> >>> > >>> >>>           TX packets:0 errors:0
dropped:0
> >>> overruns:0 carrier:0
> >>> > >>> >>>           collisions:0
txqueuelen:1000
> >>> > >>> >>>           RX bytes:0 (0.0 b) 
TX bytes:0 (0.0 b)
> >>> > >>> >>>           Interrupt:21 Base
address:0x2000
> >>> > >>> >>>
> >>> > >>> >>> eth1      Link encap:Ethernet 
HWaddr 00:16:76:B3:16:AB
> >>> > >>> >>>           inet
addr:192.168.0.101  Bcast:192.168.0.255
> >>> > >>> >>> Mask:255.255.255.0
> >>> > >>> >>>           inet6 addr: 
> fe80::216:76ff:feb3:16ab/64 Scope:Link
> >>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST
> MTU:1500  Metric:1
> >>> > >>> >>>           RX packets:198578
errors:0 dropped:0
> >>> overruns:0 frame:0
> >>> > >>> >>>           TX packets:117290
errors:0 dropped:0
> >>> overruns:0 carrier:0
> >>> > >>> >>>           collisions:0
txqueuelen:0
> >>> > >>> >>>           RX bytes:267328989
(254.9 MiB)  TX
> >>> bytes:8294632 (7.9
> >>> > >>> >>> MiB)
> >>> > >>> >>>
> >>> > >>> >>> lo        Link encap:Local
Loopback
> >>> > >>> >>>           inet addr:127.0.0.1 
Mask:255.0.0.0
> >>> > >>> >>>           inet6 addr: ::1/128
Scope:Host
> >>> > >>> >>>           UP LOOPBACK RUNNING 
MTU:16436  Metric:1
> >>> > >>> >>>           RX packets:2689
errors:0 dropped:0
> >>> overruns:0 frame:0
> >>> > >>> >>>           TX packets:2689
errors:0 dropped:0
> >>> overruns:0 carrier:0
> >>> > >>> >>>           collisions:0
txqueuelen:0
> >>> > >>> >>>           RX bytes:12510296
(11.9 MiB)  TX
> >>> bytes:12510296 (11.9
> >>> > >>> >>> MiB)
> >>> > >>> >>>
> >>> > >>> >>> peth1     Link encap:Ethernet 
HWaddr 00:16:76:B3:16:AB
> >>> > >>> >>>           inet6 addr: 
> fe80::216:76ff:feb3:16ab/64 Scope:Link
> >>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST
> MTU:1500  Metric:1
> >>> > >>> >>>           RX packets:198588
errors:0 dropped:0
> >>> overruns:0 frame:0
> >>> > >>> >>>           TX packets:117311
errors:0 dropped:0
> >>> overruns:0 carrier:0
> >>> > >>> >>>           collisions:0
txqueuelen:100
> >>> > >>> >>>           RX bytes:270906777
(258.3 MiB)  TX
> >>> bytes:8813848 (8.4
> >>> > >>> >>> MiB) Base address:0x40c0
Memory:92200000-92220000
> >>> > >>> >>>
> >>> > >>> >>> vif4.0    Link encap:Ethernet 
HWaddr FE:FF:FF:FF:FF:FF
> >>> > >>> >>>           inet6 addr:
fe80::fcff:ffff:feff:ffff/64
> >>> Scope:Link
> >>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST
> MTU:1500  Metric:1
> >>> > >>> >>>           RX packets:9 errors:0
dropped:0
> overruns:0 frame:0
> >>> > >>> >>>           TX packets:1 errors:0
dropped:6
> >>> overruns:0 carrier:0
> >>> > >>> >>>           collisions:0
txqueuelen:1
> >>> > >>> >>>           RX bytes:1068 (1.0
KiB)  TX bytes:342
> (342.0 b)
> >>> > >>> >>>
> >>> > >>> >>> virbr0    Link encap:Ethernet 
HWaddr FE:FF:FF:FF:FF:FF
> >>> > >>> >>>           inet
addr:192.168.122.1  Bcast:192.168.122.255
> >>> > >>> >>> Mask:255.255.255.0 inet6 addr:
> >>> fe80::200:ff:fe00:0/64 Scope:Link
> >>> > >>> >>>           UP BROADCAST RUNNING
MULTICAST
> MTU:1500  Metric:1
> >>> > >>> >>>           RX packets:43
errors:0 dropped:0
> >>> overruns:0 frame:0
> >>> > >>> >>>           TX packets:17
errors:0 dropped:0
> >>> overruns:0 carrier:0
> >>> > >>> >>>           collisions:0
txqueuelen:0
> >>> > >>> >>>           RX bytes:3208 (3.1
KiB)  TX
> bytes:2018 (1.9 KiB)
> >>> > >>> >>>
> >>> > >>> >>> I don''t see any xen
bridge, because that''s what I
> >>> think I need: one
> >>> > >>> >>> network card, and one xen
bridge.
> >>> > >>> >>> I found on google that I could
use the following script:
> >>> > >>> >>> #!/bin/sh
> >>> > >>> >>> dir=$(dirname "$0")
> >>> > >>> >>> "$dir/network-bridge"
"$@" vifnum=0 netdev=eth0
> >>> bridge=xenbr0
> >>> > >>> >>> "$dir/network-bridge"
"$@" vifnum=1 netdev=eth1
> >>> bridge=xenbr1
> >>> > >>> >>> "$dir/network-bridge"
"$@" vifnum=2 netdev=eth2
> >>> bridge=xenbr2
> >>> > >>> >>> And then set it into
xen-config.sxp:
> >>> > >>> >>> network-script matrix-network
> >>> > >>> >>> But it gives errors that
network-script has only
> >>> start, stop and
> >>> > >>> >>> status. The only thing that it
succeds is that I
> >>> can see a xen
> >>> > >>>
> >>> > >>> bridge.
> >>> > >>>
> >>> > >>> >>> If this would work,
doesn''t this affects other
> >>> guest domains also?
> >>> > >>> >>>
> >>> > >>> >>> My question is: How can I set a
guest dom to use
> >>> directly a network
> >>> > >>> >>> card with other ip class ?
> >>> > >>> >>>
> >>> > >>> >>> Best regards,
> >>> > >>> >>> Octav
> >>> > >>> >>>
> >>> > >>> >>>
> >>> > >>> >>>
> >>> > >>> >>>
_______________________________________________
> >>> > >>> >>> Xen-users mailing list
> >>> > >>> >>> Xen-users@lists.xensource.com
> >>> > >>> >>>
http://lists.xensource.com/xen-users
> >>> > >>> >>
> >>> > >>> >> --
> >>> > >>> >> Kobalt W.I.T.
> >>> > >>> >> Web & Information Technology
> >>> > >>> >> Brusselsesteenweg 152
> >>> > >>> >> 1850 Grimbergen
> >>> > >>> >>
> >>> > >>> >> Tel  : +32 479 339 655
> >>> > >>> >> Email: info@kobaltwit.be
> >>> > >>> >>
> >>> > >>> >>
_______________________________________________
> >>> > >>> >> Xen-users mailing list
> >>> > >>> >> Xen-users@lists.xensource.com
> >>> > >>> >>
http://lists.xensource.com/xen-users
> >>> > >>
> >>> > >> --
> >>> > >> Kobalt W.I.T.
> >>> > >> Web & Information Technology
> >>> > >> Brusselsesteenweg 152
> >>> > >> 1850 Grimbergen
> >>> > >>
> >>> > >> Tel  : +32 479 339 655
> >>> > >> Email: info@kobaltwit.be
> >>> > >>
> >>> > >> _______________________________________________
> >>> > >> Xen-users mailing list
> >>> > >> Xen-users@lists.xensource.com
> >>> > >> http://lists.xensource.com/xen-users
> >>> >
> >>> > _______________________________________________
> >>> > Xen-users mailing list
> >>> > Xen-users@lists.xensource.com
> >>> > http://lists.xensource.com/xen-users
> >>>
> >>> --
> >>> Kobalt W.I.T.
> >>> Web & Information Technology
> >>> Brusselsesteenweg 152
> >>> 1850 Grimbergen
> >>>
> >>> Tel  : +32 479 339 655
> >>> Email: info@kobaltwit.be
> >>>
> >>> _______________________________________________
> >>> Xen-users mailing list
> >>> Xen-users@lists.xensource.com
> >>> http://lists.xensource.com/xen-users
> >>>
> >>>
> >>>
> >>
> >>
> >>
> >> _______________________________________________
> >> Xen-users mailing list
> >> Xen-users@lists.xensource.com
> >> http://lists.xensource.com/xen-users
> >>
> 
> 
> 
> 
> 


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

Petersson, Mats

2007-Jun-05 14:12 UTC

head link

RE: [Xen-users] Xen a couple of questions

> -----Original Message-----
> From: Octavian Teodorescu [mailto:octav@inetsolutions.ro] 
> Sent: 05 June 2007 15:13
> To: Petersson, Mats
> Cc: xen-users@lists.xensource.com
> Subject: RE: [Xen-users] Xen a couple of questions
> 
> Tried to create the ramdisk with blkfront module:
> mkinitrd --with=blkfront initrd-test.img 2.6.20-2925.9.fc7xen
> ... but I get the same error.
What does you config file look like? 

--
Mats> 
> 
> >>
> >>
> >>> -----Original Message-----
> >>> From: Octavian Teodorescu [mailto:octav@inetsolutions.ro]
> >>> Sent: 05 June 2007 13:50
> >>> To: Petersson, Mats
> >>> Cc: xen-users@lists.xensource.com
> >>> Subject: RE: [Xen-users] Xen a couple of questions
> >>>
> >>> Yap, I have the ramdisk of dom0 also. Without the ramdisk 
> set up, I''ve
> >>> seen that it can''t mount logical volumes as
partitions (with
> >>> any kind of
> >>> kernel).
> >>
> >> Yes, but you need one that uses blkfront driver if you are 
> using it in
> >> domU''s. You can technically create the same one for both 
> DomU and Dom0,
> >> but I think it''s easier/better to have separate ones, as 
> at least you
> >> don''t risk breaking the working Dom0 one trying to set
the
> DomU one up.
> >>
> >> --
> >> Mats
> >>>
> >>> Best regards.
> >>>
> >>> >>
> >>> >>
> >>> >>> -----Original Message-----
> >>> >>> From: xen-users-bounces@lists.xensource.com
> >>> >>> [mailto:xen-users-bounces@lists.xensource.com] On
Behalf Of
> >>> >>> Geert Janssens
> >>> >>> Sent: 05 June 2007 13:26
> >>> >>> To: xen-users@lists.xensource.com
> >>> >>> Cc: Octavian Teodorescu
> >>> >>> Subject: Re: [Xen-users] Xen a couple of
questions
> >>> >>>
> >>> >>> I am sorry, this is way over my head.
I''m just a (relatively
> >>> >>> new) user of xen
> >>> >>> myself. I hope someone else can help you here.
> >>> >>>
> >>> >>> Regards,
> >>> >>>
> >>> >>> Geert
> >>> >>>
> >>> >>> On Tuesday 5 June 2007 14:19, Octavian Teodorescu
wrote:
> >>> >>> > I had stoped the virtual domain and modified
in the config
> >>> >>> file to use the
> >>> >>> > dom0 kernel.
> >>> >>> > Here are the last lines of the guest boot
where you can see
> >>> >>> the error:
> >>> >>> > "SCSI subsystem initialized
> >>> >>> > device-mapper: ioctl: 4.11.0-ioctl
(2006-10-12) initialised:
> >>> >>> > dm-devel@redhat.com
> >>> >>> > Kernel panic - not syncing: Attempted to
kill init!"
> >>> >>> >
> >>> >>> > But if I use the old kernelU i have which is
actually for
> >>> >>> Fedorac core 5
> >>> >>> > then everthing it''s ok.
> >>> >>
> >>> >> Do you have an "ramdisk" entry in your PV
domain config?
> >>> You probably
> >>> >> need one.
> >>> >>
> >>> >> --
> >>> >> Mats
> >>> >>> >
> >>> >>> > >> On Tuesday 5 June 2007 13:11, you
wrote:
> >>> >>> > >>> Thanks a lot, I appreciate your
help.
> >>> >>> > >>> 1. I tried with the same
configuration and with the
> >>> >>> kernel of dom0, but
> >>> >>> > >>> I
> >>> >>> > >>> receive a lot of errors, on
both Fedora Core 7 and
> >>> >>> Centos 5 (systems
> >>> >>> > >>> with
> >>> >>> > >>> which I''ve tried xen).
> >>> >>> > >>
> >>> >>> > >> Hmm, I don''t know what
exactly you tried and
> what is failing.
> >>> >>> > >>
> >>> >>> > >> I simply used virt-manager to
create a CentOS 5 guest on
> >>> >>> my CentOS 5
> >>> >>> > >> dom0. I
> >>> >>> > >> followed (more or less) the
guidelines that come
> >>> with CentOS 5''s
> >>> >>> > >> release:
> >>> http://www.centos.org/docs/5/html/Virtualization-en-US/
> >>> >>> > >>
> >>> >>> > >> I don''t remember having
particular difficulty with this.
> >>> >>> The only caveat
> >>> >>> > >> I remember was that Anaconda
insists on a block device
> >>> >>> that can hold a
> >>> >>> > >> partition map. So while installing,
you can''t provide
> >>> >>> the guest with
> >>> >>> > >> separate
> >>> >>> > >> partitions, because it will treat
these separate
> >>> >>> partitions as complete
> >>> >>> > >> diskss that have to be partitioned
still.
> >>> >>> > >>
> >>> >>> > >> Regards,
> >>> >>> > >>
> >>> >>> > >> Geert
> >>> >>> > >>
> >>> >>> > >>> 2.Yap that''s a bridge
interface. I''ll look into
> >>> your mails about
> >>> >>> > >>> advance bridging, thanks.
> >>> >>> > >>>
> >>> >>> > >>> Best regards.
> >>> >>> > >>>
> >>> >>> > >>> >> 1. On CentOS 5, Redhat
Enterprise 5 and Fedora core
> >>> >>> 6 and up, the
> >>> >>> > >>> >> xen kernel
> >>> >>> > >>> >> can be used for both
dom0 and domU. There is no need
> >>> >>> anymore for two
> >>> >>> > >>> >> kernels.
> >>> >>> > >>> >>
> >>> >>> > >>> >> 2. I don''t
know the complete answer to your second
> >>> >>> question. From
> >>> >>> > >>>
> >>> >>> > >>> your
> >>> >>> > >>>
> >>> >>> > >>> >> ifconfig output, it
looks as if Fedora Core 7
> >>> >>> creates a virbr0
> >>> >>> > >>>
> >>> >>> > >>> interface
> >>> >>> > >>>
> >>> >>> > >>> >> instead of a xenbr0
interface. You could check if
> >>> >>> this is really a
> >>> >>> > >>> >> bridge with the
command "brctl show". You probably
> >>> >>> have to execute
> >>> >>> > >>>
> >>> >>> > >>> this
> >>> >>> > >>>
> >>> >>> > >>> >> as root.
> >>> >>> > >>> >>
> >>> >>> > >>> >> Then if I understand
your question correctly, you
> >>> >>> are trying to
> >>> >>> > >>> >> setup
> >>> >>> > >>>
> >>> >>> > >>> a
> >>> >>> > >>>
> >>> >>> > >>> >> xen
> >>> >>> > >>> >> guest domain to act as
a
> >>> >>> firewall/router/gateway/whatever for your
> >>> >>> > >>>
> >>> >>> > >>> lan.
> >>> >>> > >>>
> >>> >>> > >>> >> So I assume you only
want this guest domain to use
> >>> >>> the external
> >>> >>> > >>>
> >>> >>> > >>> network
> >>> >>> > >>>
> >>> >>> > >>> >> card
> >>> >>> > >>> >> (your eth0). There are
two ways of
> accomplishing this:
> >>> >>> > >>> >> * either use PCI
passthrough so that your dom0 won''t
> >>> >>> see eth0, but
> >>> >>> > >>> >> instead it''s
passed to your guest system (search for
> >>> >>> pciback on
> >>> >>> > >>>
> >>> >>> > >>> Google
> >>> >>> > >>>
> >>> >>> > >>> >> for more info).
Unfortunatly, I didn''t manage to set
> >>> >>> this up in my
> >>> >>> > >>> >> particular case, so
> >>> >>> > >>> >> I used the seconde
option:
> >>> >>> > >>> >> * create two
xenbridges, one for your external
> >>> >>> network interface,
> >>> >>> > >>> >> and one for
> >>> >>> > >>> >> your internal network
interface. Then configure dom0
> >>> >>> such that it
> >>> >>> > >>>
> >>> >>> > >>> isn''t
> >>> >>> > >>>
> >>> >>> > >>> >> allowed to use the
bridge for the external
> >>> >>> interface. You can do
> >>> >>> > >>> >> this
> >>> >>> > >>>
> >>> >>> > >>> by
> >>> >>> > >>>
> >>> >>> > >>> >> either disabling the
virtual interface in dom0
> >>> >>> (which will be called
> >>> >>> > >>> >> eth0) or
> >>> >>> > >>> >> by setting some
firewall rules in dom0, or both.
> >>> >>> > >>> >> You can search this
list for one of my earlier
> >>> mails, where I
> >>> >>> > >>> >> explain
> >>> >>> > >>>
> >>> >>> > >>> my
> >>> >>> > >>>
> >>> >>> > >>> >> configuration (on
CentOS 5). It''s titled "advanced
> >>> >>> bridging..." and
> >>> >>> > >>> >> dated May
> >>> >>> > >>> >> 16th, 2007.
> >>> >>> > >>> >>
> >>> >>> > >>> >> Hopefully this will
help you along the way.
> >>> >>> > >>> >>
> >>> >>> > >>> >> Cheers,
> >>> >>> > >>> >>
> >>> >>> > >>> >> Geert
> >>> >>> > >>> >>
> >>> >>> > >>> >> On Tuesday 5 June 2007
10:34, Octavian
> Teodorescu wrote:
> >>> >>> > >>> >>> Hi guys,
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> 1. Regarding
Centos and Fedora core 7 compared with
> >>> >>> fedora core 5.
> >>> >>> > >>>
> >>> >>> > >>> I''ve
> >>> >>> > >>>
> >>> >>> > >>> >>> seen that on
fedora core 5 when you want to install
> >>> >>> xen you have to
> >>> >>> > >>> >>> install the
following packages: xen, kernel-xen0
> >>> >>> and kernel-xenU
> >>> >>> > >>> >>> (of course with
the dependencies needed). But on
> >>> >>> Centos, FC7 and I
> >>> >>> > >>> >>> think redhat
versions, you only have to install xen
> >>> >>> and kernel-xen,
> >>> >>> > >>> >>> you
> >>> >>> > >>>
> >>> >>> > >>> don''t
> >>> >>> > >>>
> >>> >>> > >>> >>> have any kernel
for the guest system. In my case I
> >>> >>> could only start
> >>> >>> > >>>
> >>> >>> > >>> a
> >>> >>> > >>>
> >>> >>> > >>> >>> xen
> >>> >>> > >>> >>> guest (on FC7)
with an older kernel-xenU installed
> >>> >>> from FC version
> >>> >>> > >>>
> >>> >>> > >>> 5.
> >>> >>> > >>>
> >>> >>> > >>> >>> My question is:
Why does the newer releases of
> >>> >>> linux has xen kernel
> >>> >>> > >>> >>> prebuilt but just
for dom0, not for the guest
> >>> >>> systems, and you
> >>> >>> > >>> >>> can''t
even
> >>> >>> > >>> >>> find a domU kernel
special for those systems?
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> 2.My network
topology in my home is like this:
> >>> >>> > >>> >>> --------
> >>> >>> > >>> >>> -router-
> >>> >>> > >>> >>> --------
> >>> >>> > >>> >>>
> >>> >>> > >>> >>>
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> -----------       
------------
> >>> >>> > >>> >>> -linux xen-   ----
-other 2 pc-
> >>> >>> > >>> >>> -----------       
------------
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> The linux xen
machine has two network interfaces
> >>> >>> and xen installed.
> >>> >>> > >>> >>> I want: -  one
windows machine virtualized
> >>> >>> > >>> >>>         -  one
linux machine for which I want to
> >>> >>> have a public
> >>> >>> > >>> >>> ipaddress (to put
the ip in DMS on the router) and
> >>> >>> I want it to use
> >>> >>> > >>> >>> eth0 (so in this
case the traffic can not be
> >>> >>> sniffed by other guest
> >>> >>> > >>> >>> systems or dom0).
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> ifconfig -a (on
dom0) it shows like this:
> >>> >>> > >>> >>> eth0      Link
encap:Ethernet  HWaddr
> 00:00:E8:76:E2:4D
> >>> >>> > >>> >>>           UP
BROADCAST MULTICAST  MTU:1500  Metric:1
> >>> >>> > >>> >>>           RX
packets:0 errors:0 dropped:0
> >>> overruns:0 frame:0
> >>> >>> > >>> >>>           TX
packets:0 errors:0 dropped:0
> >>> >>> overruns:0 carrier:0
> >>> >>> > >>> >>>          
collisions:0 txqueuelen:1000
> >>> >>> > >>> >>>           RX
bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
> >>> >>> > >>> >>>          
Interrupt:21 Base address:0x2000
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> eth1      Link
encap:Ethernet  HWaddr
> 00:16:76:B3:16:AB
> >>> >>> > >>> >>>           inet
addr:192.168.0.101
> Bcast:192.168.0.255
> >>> >>> > >>> >>> Mask:255.255.255.0
> >>> >>> > >>> >>>           inet6
addr:
> >>> fe80::216:76ff:feb3:16ab/64 Scope:Link
> >>> >>> > >>> >>>           UP
BROADCAST RUNNING MULTICAST
> >>> MTU:1500  Metric:1
> >>> >>> > >>> >>>           RX
packets:198578 errors:0 dropped:0
> >>> >>> overruns:0 frame:0
> >>> >>> > >>> >>>           TX
packets:117290 errors:0 dropped:0
> >>> >>> overruns:0 carrier:0
> >>> >>> > >>> >>>          
collisions:0 txqueuelen:0
> >>> >>> > >>> >>>           RX
bytes:267328989 (254.9 MiB)  TX
> >>> >>> bytes:8294632 (7.9
> >>> >>> > >>> >>> MiB)
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> lo        Link
encap:Local Loopback
> >>> >>> > >>> >>>           inet
addr:127.0.0.1  Mask:255.0.0.0
> >>> >>> > >>> >>>           inet6
addr: ::1/128 Scope:Host
> >>> >>> > >>> >>>           UP
LOOPBACK RUNNING  MTU:16436  Metric:1
> >>> >>> > >>> >>>           RX
packets:2689 errors:0 dropped:0
> >>> >>> overruns:0 frame:0
> >>> >>> > >>> >>>           TX
packets:2689 errors:0 dropped:0
> >>> >>> overruns:0 carrier:0
> >>> >>> > >>> >>>          
collisions:0 txqueuelen:0
> >>> >>> > >>> >>>           RX
bytes:12510296 (11.9 MiB)  TX
> >>> >>> bytes:12510296 (11.9
> >>> >>> > >>> >>> MiB)
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> peth1     Link
encap:Ethernet  HWaddr
> 00:16:76:B3:16:AB
> >>> >>> > >>> >>>           inet6
addr:
> >>> fe80::216:76ff:feb3:16ab/64 Scope:Link
> >>> >>> > >>> >>>           UP
BROADCAST RUNNING MULTICAST
> >>> MTU:1500  Metric:1
> >>> >>> > >>> >>>           RX
packets:198588 errors:0 dropped:0
> >>> >>> overruns:0 frame:0
> >>> >>> > >>> >>>           TX
packets:117311 errors:0 dropped:0
> >>> >>> overruns:0 carrier:0
> >>> >>> > >>> >>>          
collisions:0 txqueuelen:100
> >>> >>> > >>> >>>           RX
bytes:270906777 (258.3 MiB)  TX
> >>> >>> bytes:8813848 (8.4
> >>> >>> > >>> >>> MiB) Base
address:0x40c0 Memory:92200000-92220000
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> vif4.0    Link
encap:Ethernet  HWaddr
> FE:FF:FF:FF:FF:FF
> >>> >>> > >>> >>>           inet6
addr: fe80::fcff:ffff:feff:ffff/64
> >>> >>> Scope:Link
> >>> >>> > >>> >>>           UP
BROADCAST RUNNING MULTICAST
> >>> MTU:1500  Metric:1
> >>> >>> > >>> >>>           RX
packets:9 errors:0 dropped:0
> >>> overruns:0 frame:0
> >>> >>> > >>> >>>           TX
packets:1 errors:0 dropped:6
> >>> >>> overruns:0 carrier:0
> >>> >>> > >>> >>>          
collisions:0 txqueuelen:1
> >>> >>> > >>> >>>           RX
bytes:1068 (1.0 KiB)  TX bytes:342
> >>> (342.0 b)
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> virbr0    Link
encap:Ethernet  HWaddr
> FE:FF:FF:FF:FF:FF
> >>> >>> > >>> >>>           inet
addr:192.168.122.1
> Bcast:192.168.122.255
> >>> >>> > >>> >>> Mask:255.255.255.0
inet6 addr:
> >>> >>> fe80::200:ff:fe00:0/64 Scope:Link
> >>> >>> > >>> >>>           UP
BROADCAST RUNNING MULTICAST
> >>> MTU:1500  Metric:1
> >>> >>> > >>> >>>           RX
packets:43 errors:0 dropped:0
> >>> >>> overruns:0 frame:0
> >>> >>> > >>> >>>           TX
packets:17 errors:0 dropped:0
> >>> >>> overruns:0 carrier:0
> >>> >>> > >>> >>>          
collisions:0 txqueuelen:0
> >>> >>> > >>> >>>           RX
bytes:3208 (3.1 KiB)  TX
> >>> bytes:2018 (1.9 KiB)
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> I don''t
see any xen bridge, because that''s what I
> >>> >>> think I need: one
> >>> >>> > >>> >>> network card, and
one xen bridge.
> >>> >>> > >>> >>> I found on google
that I could use the
> following script:
> >>> >>> > >>> >>> #!/bin/sh
> >>> >>> > >>> >>> dir=$(dirname
"$0")
> >>> >>> > >>> >>>
"$dir/network-bridge" "$@" vifnum=0 netdev=eth0
> >>> >>> bridge=xenbr0
> >>> >>> > >>> >>>
"$dir/network-bridge" "$@" vifnum=1 netdev=eth1
> >>> >>> bridge=xenbr1
> >>> >>> > >>> >>>
"$dir/network-bridge" "$@" vifnum=2 netdev=eth2
> >>> >>> bridge=xenbr2
> >>> >>> > >>> >>> And then set it
into xen-config.sxp:
> >>> >>> > >>> >>> network-script
matrix-network
> >>> >>> > >>> >>> But it gives
errors that network-script has only
> >>> >>> start, stop and
> >>> >>> > >>> >>> status. The only
thing that it succeds is that I
> >>> >>> can see a xen
> >>> >>> > >>>
> >>> >>> > >>> bridge.
> >>> >>> > >>>
> >>> >>> > >>> >>> If this would
work, doesn''t this affects other
> >>> >>> guest domains also?
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> My question is:
How can I set a guest dom to use
> >>> >>> directly a network
> >>> >>> > >>> >>> card with other ip
class ?
> >>> >>> > >>> >>>
> >>> >>> > >>> >>> Best regards,
> >>> >>> > >>> >>> Octav
> >>> >>> > >>> >>>
> >>> >>> > >>> >>>
> >>> >>> > >>> >>>
> >>> >>> > >>> >>>
_______________________________________________
> >>> >>> > >>> >>> Xen-users mailing
list
> >>> >>> > >>> >>>
Xen-users@lists.xensource.com
> >>> >>> > >>> >>>
http://lists.xensource.com/xen-users
> >>> >>> > >>> >>
> >>> >>> > >>> >> --
> >>> >>> > >>> >> Kobalt W.I.T.
> >>> >>> > >>> >> Web & Information
Technology
> >>> >>> > >>> >> Brusselsesteenweg 152
> >>> >>> > >>> >> 1850 Grimbergen
> >>> >>> > >>> >>
> >>> >>> > >>> >> Tel  : +32 479 339 655
> >>> >>> > >>> >> Email:
info@kobaltwit.be
> >>> >>> > >>> >>
> >>> >>> > >>> >>
_______________________________________________
> >>> >>> > >>> >> Xen-users mailing list
> >>> >>> > >>> >>
Xen-users@lists.xensource.com
> >>> >>> > >>> >>
http://lists.xensource.com/xen-users
> >>> >>> > >>
> >>> >>> > >> --
> >>> >>> > >> Kobalt W.I.T.
> >>> >>> > >> Web & Information Technology
> >>> >>> > >> Brusselsesteenweg 152
> >>> >>> > >> 1850 Grimbergen
> >>> >>> > >>
> >>> >>> > >> Tel  : +32 479 339 655
> >>> >>> > >> Email: info@kobaltwit.be
> >>> >>> > >>
> >>> >>> > >>
_______________________________________________
> >>> >>> > >> Xen-users mailing list
> >>> >>> > >> Xen-users@lists.xensource.com
> >>> >>> > >>
http://lists.xensource.com/xen-users
> >>> >>> >
> >>> >>> >
_______________________________________________
> >>> >>> > Xen-users mailing list
> >>> >>> > Xen-users@lists.xensource.com
> >>> >>> > http://lists.xensource.com/xen-users
> >>> >>>
> >>> >>> --
> >>> >>> Kobalt W.I.T.
> >>> >>> Web & Information Technology
> >>> >>> Brusselsesteenweg 152
> >>> >>> 1850 Grimbergen
> >>> >>>
> >>> >>> Tel  : +32 479 339 655
> >>> >>> Email: info@kobaltwit.be
> >>> >>>
> >>> >>> _______________________________________________
> >>> >>> Xen-users mailing list
> >>> >>> Xen-users@lists.xensource.com
> >>> >>> http://lists.xensource.com/xen-users
> >>> >>>
> >>> >>>
> >>> >>>
> >>> >>
> >>> >>
> >>> >>
> >>> >> _______________________________________________
> >>> >> Xen-users mailing list
> >>> >> Xen-users@lists.xensource.com
> >>> >> http://lists.xensource.com/xen-users
> >>> >>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>
> >>
> >>
> 
> 
> 
> 
> 


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

Octavian Teodorescu

2007-Jun-05 14:13 UTC

head link

RE: [Xen-users] Xen a couple of questions

Tried to create the ramdisk with blkfront module:
mkinitrd --with=blkfront initrd-test.img 2.6.20-2925.9.fc7xen
... but I get the same error.

>>
>>
>>> -----Original Message-----
>>> From: Octavian Teodorescu [mailto:octav@inetsolutions.ro]
>>> Sent: 05 June 2007 13:50
>>> To: Petersson, Mats
>>> Cc: xen-users@lists.xensource.com
>>> Subject: RE: [Xen-users] Xen a couple of questions
>>>
>>> Yap, I have the ramdisk of dom0 also. Without the ramdisk set up,
I''ve
>>> seen that it can''t mount logical volumes as partitions
(with
>>> any kind of
>>> kernel).
>>
>> Yes, but you need one that uses blkfront driver if you are using it in
>> domU''s. You can technically create the same one for both DomU
and Dom0,
>> but I think it''s easier/better to have separate ones, as at
least you
>> don''t risk breaking the working Dom0 one trying to set the
DomU one up.
>>
>> --
>> Mats
>>>
>>> Best regards.
>>>
>>> >>
>>> >>
>>> >>> -----Original Message-----
>>> >>> From: xen-users-bounces@lists.xensource.com
>>> >>> [mailto:xen-users-bounces@lists.xensource.com] On
Behalf Of
>>> >>> Geert Janssens
>>> >>> Sent: 05 June 2007 13:26
>>> >>> To: xen-users@lists.xensource.com
>>> >>> Cc: Octavian Teodorescu
>>> >>> Subject: Re: [Xen-users] Xen a couple of questions
>>> >>>
>>> >>> I am sorry, this is way over my head. I''m
just a (relatively
>>> >>> new) user of xen
>>> >>> myself. I hope someone else can help you here.
>>> >>>
>>> >>> Regards,
>>> >>>
>>> >>> Geert
>>> >>>
>>> >>> On Tuesday 5 June 2007 14:19, Octavian Teodorescu
wrote:
>>> >>> > I had stoped the virtual domain and modified in
the config
>>> >>> file to use the
>>> >>> > dom0 kernel.
>>> >>> > Here are the last lines of the guest boot where
you can see
>>> >>> the error:
>>> >>> > "SCSI subsystem initialized
>>> >>> > device-mapper: ioctl: 4.11.0-ioctl (2006-10-12)
initialised:
>>> >>> > dm-devel@redhat.com
>>> >>> > Kernel panic - not syncing: Attempted to kill
init!"
>>> >>> >
>>> >>> > But if I use the old kernelU i have which is
actually for
>>> >>> Fedorac core 5
>>> >>> > then everthing it''s ok.
>>> >>
>>> >> Do you have an "ramdisk" entry in your PV domain
config?
>>> You probably
>>> >> need one.
>>> >>
>>> >> --
>>> >> Mats
>>> >>> >
>>> >>> > >> On Tuesday 5 June 2007 13:11, you wrote:
>>> >>> > >>> Thanks a lot, I appreciate your
help.
>>> >>> > >>> 1. I tried with the same
configuration and with the
>>> >>> kernel of dom0, but
>>> >>> > >>> I
>>> >>> > >>> receive a lot of errors, on both
Fedora Core 7 and
>>> >>> Centos 5 (systems
>>> >>> > >>> with
>>> >>> > >>> which I''ve tried xen).
>>> >>> > >>
>>> >>> > >> Hmm, I don''t know what exactly
you tried and what is failing.
>>> >>> > >>
>>> >>> > >> I simply used virt-manager to create a
CentOS 5 guest on
>>> >>> my CentOS 5
>>> >>> > >> dom0. I
>>> >>> > >> followed (more or less) the guidelines
that come
>>> with CentOS 5''s
>>> >>> > >> release:
>>> http://www.centos.org/docs/5/html/Virtualization-en-US/
>>> >>> > >>
>>> >>> > >> I don''t remember having
particular difficulty with this.
>>> >>> The only caveat
>>> >>> > >> I remember was that Anaconda insists on
a block device
>>> >>> that can hold a
>>> >>> > >> partition map. So while installing, you
can''t provide
>>> >>> the guest with
>>> >>> > >> separate
>>> >>> > >> partitions, because it will treat these
separate
>>> >>> partitions as complete
>>> >>> > >> diskss that have to be partitioned
still.
>>> >>> > >>
>>> >>> > >> Regards,
>>> >>> > >>
>>> >>> > >> Geert
>>> >>> > >>
>>> >>> > >>> 2.Yap that''s a bridge
interface. I''ll look into
>>> your mails about
>>> >>> > >>> advance bridging, thanks.
>>> >>> > >>>
>>> >>> > >>> Best regards.
>>> >>> > >>>
>>> >>> > >>> >> 1. On CentOS 5, Redhat
Enterprise 5 and Fedora core
>>> >>> 6 and up, the
>>> >>> > >>> >> xen kernel
>>> >>> > >>> >> can be used for both dom0
and domU. There is no need
>>> >>> anymore for two
>>> >>> > >>> >> kernels.
>>> >>> > >>> >>
>>> >>> > >>> >> 2. I don''t know
the complete answer to your second
>>> >>> question. From
>>> >>> > >>>
>>> >>> > >>> your
>>> >>> > >>>
>>> >>> > >>> >> ifconfig output, it looks
as if Fedora Core 7
>>> >>> creates a virbr0
>>> >>> > >>>
>>> >>> > >>> interface
>>> >>> > >>>
>>> >>> > >>> >> instead of a xenbr0
interface. You could check if
>>> >>> this is really a
>>> >>> > >>> >> bridge with the command
"brctl show". You probably
>>> >>> have to execute
>>> >>> > >>>
>>> >>> > >>> this
>>> >>> > >>>
>>> >>> > >>> >> as root.
>>> >>> > >>> >>
>>> >>> > >>> >> Then if I understand your
question correctly, you
>>> >>> are trying to
>>> >>> > >>> >> setup
>>> >>> > >>>
>>> >>> > >>> a
>>> >>> > >>>
>>> >>> > >>> >> xen
>>> >>> > >>> >> guest domain to act as a
>>> >>> firewall/router/gateway/whatever for your
>>> >>> > >>>
>>> >>> > >>> lan.
>>> >>> > >>>
>>> >>> > >>> >> So I assume you only want
this guest domain to use
>>> >>> the external
>>> >>> > >>>
>>> >>> > >>> network
>>> >>> > >>>
>>> >>> > >>> >> card
>>> >>> > >>> >> (your eth0). There are two
ways of accomplishing this:
>>> >>> > >>> >> * either use PCI
passthrough so that your dom0 won''t
>>> >>> see eth0, but
>>> >>> > >>> >> instead it''s
passed to your guest system (search for
>>> >>> pciback on
>>> >>> > >>>
>>> >>> > >>> Google
>>> >>> > >>>
>>> >>> > >>> >> for more info).
Unfortunatly, I didn''t manage to set
>>> >>> this up in my
>>> >>> > >>> >> particular case, so
>>> >>> > >>> >> I used the seconde option:
>>> >>> > >>> >> * create two xenbridges,
one for your external
>>> >>> network interface,
>>> >>> > >>> >> and one for
>>> >>> > >>> >> your internal network
interface. Then configure dom0
>>> >>> such that it
>>> >>> > >>>
>>> >>> > >>> isn''t
>>> >>> > >>>
>>> >>> > >>> >> allowed to use the bridge
for the external
>>> >>> interface. You can do
>>> >>> > >>> >> this
>>> >>> > >>>
>>> >>> > >>> by
>>> >>> > >>>
>>> >>> > >>> >> either disabling the
virtual interface in dom0
>>> >>> (which will be called
>>> >>> > >>> >> eth0) or
>>> >>> > >>> >> by setting some firewall
rules in dom0, or both.
>>> >>> > >>> >> You can search this list
for one of my earlier
>>> mails, where I
>>> >>> > >>> >> explain
>>> >>> > >>>
>>> >>> > >>> my
>>> >>> > >>>
>>> >>> > >>> >> configuration (on CentOS
5). It''s titled "advanced
>>> >>> bridging..." and
>>> >>> > >>> >> dated May
>>> >>> > >>> >> 16th, 2007.
>>> >>> > >>> >>
>>> >>> > >>> >> Hopefully this will help
you along the way.
>>> >>> > >>> >>
>>> >>> > >>> >> Cheers,
>>> >>> > >>> >>
>>> >>> > >>> >> Geert
>>> >>> > >>> >>
>>> >>> > >>> >> On Tuesday 5 June 2007
10:34, Octavian Teodorescu wrote:
>>> >>> > >>> >>> Hi guys,
>>> >>> > >>> >>>
>>> >>> > >>> >>> 1. Regarding Centos and
Fedora core 7 compared with
>>> >>> fedora core 5.
>>> >>> > >>>
>>> >>> > >>> I''ve
>>> >>> > >>>
>>> >>> > >>> >>> seen that on fedora
core 5 when you want to install
>>> >>> xen you have to
>>> >>> > >>> >>> install the following
packages: xen, kernel-xen0
>>> >>> and kernel-xenU
>>> >>> > >>> >>> (of course with the
dependencies needed). But on
>>> >>> Centos, FC7 and I
>>> >>> > >>> >>> think redhat versions,
you only have to install xen
>>> >>> and kernel-xen,
>>> >>> > >>> >>> you
>>> >>> > >>>
>>> >>> > >>> don''t
>>> >>> > >>>
>>> >>> > >>> >>> have any kernel for the
guest system. In my case I
>>> >>> could only start
>>> >>> > >>>
>>> >>> > >>> a
>>> >>> > >>>
>>> >>> > >>> >>> xen
>>> >>> > >>> >>> guest (on FC7) with an
older kernel-xenU installed
>>> >>> from FC version
>>> >>> > >>>
>>> >>> > >>> 5.
>>> >>> > >>>
>>> >>> > >>> >>> My question is: Why
does the newer releases of
>>> >>> linux has xen kernel
>>> >>> > >>> >>> prebuilt but just for
dom0, not for the guest
>>> >>> systems, and you
>>> >>> > >>> >>> can''t even
>>> >>> > >>> >>> find a domU kernel
special for those systems?
>>> >>> > >>> >>>
>>> >>> > >>> >>> 2.My network topology
in my home is like this:
>>> >>> > >>> >>> --------
>>> >>> > >>> >>> -router-
>>> >>> > >>> >>> --------
>>> >>> > >>> >>>
>>> >>> > >>> >>>
>>> >>> > >>> >>>
>>> >>> > >>> >>> -----------        
------------
>>> >>> > >>> >>> -linux xen-   ---- 
-other 2 pc-
>>> >>> > >>> >>> -----------        
------------
>>> >>> > >>> >>>
>>> >>> > >>> >>> The linux xen machine
has two network interfaces
>>> >>> and xen installed.
>>> >>> > >>> >>> I want: -  one windows
machine virtualized
>>> >>> > >>> >>>         -  one linux
machine for which I want to
>>> >>> have a public
>>> >>> > >>> >>> ipaddress (to put the
ip in DMS on the router) and
>>> >>> I want it to use
>>> >>> > >>> >>> eth0 (so in this case
the traffic can not be
>>> >>> sniffed by other guest
>>> >>> > >>> >>> systems or dom0).
>>> >>> > >>> >>>
>>> >>> > >>> >>> ifconfig -a (on dom0)
it shows like this:
>>> >>> > >>> >>> eth0      Link
encap:Ethernet  HWaddr 00:00:E8:76:E2:4D
>>> >>> > >>> >>>           UP BROADCAST
MULTICAST  MTU:1500  Metric:1
>>> >>> > >>> >>>           RX packets:0
errors:0 dropped:0
>>> overruns:0 frame:0
>>> >>> > >>> >>>           TX packets:0
errors:0 dropped:0
>>> >>> overruns:0 carrier:0
>>> >>> > >>> >>>           collisions:0
txqueuelen:1000
>>> >>> > >>> >>>           RX bytes:0
(0.0 b)  TX bytes:0 (0.0 b)
>>> >>> > >>> >>>           Interrupt:21
Base address:0x2000
>>> >>> > >>> >>>
>>> >>> > >>> >>> eth1      Link
encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>>> >>> > >>> >>>           inet
addr:192.168.0.101  Bcast:192.168.0.255
>>> >>> > >>> >>> Mask:255.255.255.0
>>> >>> > >>> >>>           inet6 addr:
>>> fe80::216:76ff:feb3:16ab/64 Scope:Link
>>> >>> > >>> >>>           UP BROADCAST
RUNNING MULTICAST
>>> MTU:1500  Metric:1
>>> >>> > >>> >>>           RX
packets:198578 errors:0 dropped:0
>>> >>> overruns:0 frame:0
>>> >>> > >>> >>>           TX
packets:117290 errors:0 dropped:0
>>> >>> overruns:0 carrier:0
>>> >>> > >>> >>>           collisions:0
txqueuelen:0
>>> >>> > >>> >>>           RX
bytes:267328989 (254.9 MiB)  TX
>>> >>> bytes:8294632 (7.9
>>> >>> > >>> >>> MiB)
>>> >>> > >>> >>>
>>> >>> > >>> >>> lo        Link
encap:Local Loopback
>>> >>> > >>> >>>           inet
addr:127.0.0.1  Mask:255.0.0.0
>>> >>> > >>> >>>           inet6 addr:
::1/128 Scope:Host
>>> >>> > >>> >>>           UP LOOPBACK
RUNNING  MTU:16436  Metric:1
>>> >>> > >>> >>>           RX
packets:2689 errors:0 dropped:0
>>> >>> overruns:0 frame:0
>>> >>> > >>> >>>           TX
packets:2689 errors:0 dropped:0
>>> >>> overruns:0 carrier:0
>>> >>> > >>> >>>           collisions:0
txqueuelen:0
>>> >>> > >>> >>>           RX
bytes:12510296 (11.9 MiB)  TX
>>> >>> bytes:12510296 (11.9
>>> >>> > >>> >>> MiB)
>>> >>> > >>> >>>
>>> >>> > >>> >>> peth1     Link
encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>>> >>> > >>> >>>           inet6 addr:
>>> fe80::216:76ff:feb3:16ab/64 Scope:Link
>>> >>> > >>> >>>           UP BROADCAST
RUNNING MULTICAST
>>> MTU:1500  Metric:1
>>> >>> > >>> >>>           RX
packets:198588 errors:0 dropped:0
>>> >>> overruns:0 frame:0
>>> >>> > >>> >>>           TX
packets:117311 errors:0 dropped:0
>>> >>> overruns:0 carrier:0
>>> >>> > >>> >>>           collisions:0
txqueuelen:100
>>> >>> > >>> >>>           RX
bytes:270906777 (258.3 MiB)  TX
>>> >>> bytes:8813848 (8.4
>>> >>> > >>> >>> MiB) Base
address:0x40c0 Memory:92200000-92220000
>>> >>> > >>> >>>
>>> >>> > >>> >>> vif4.0    Link
encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>>> >>> > >>> >>>           inet6 addr:
fe80::fcff:ffff:feff:ffff/64
>>> >>> Scope:Link
>>> >>> > >>> >>>           UP BROADCAST
RUNNING MULTICAST
>>> MTU:1500  Metric:1
>>> >>> > >>> >>>           RX packets:9
errors:0 dropped:0
>>> overruns:0 frame:0
>>> >>> > >>> >>>           TX packets:1
errors:0 dropped:6
>>> >>> overruns:0 carrier:0
>>> >>> > >>> >>>           collisions:0
txqueuelen:1
>>> >>> > >>> >>>           RX bytes:1068
(1.0 KiB)  TX bytes:342
>>> (342.0 b)
>>> >>> > >>> >>>
>>> >>> > >>> >>> virbr0    Link
encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>>> >>> > >>> >>>           inet
addr:192.168.122.1  Bcast:192.168.122.255
>>> >>> > >>> >>> Mask:255.255.255.0
inet6 addr:
>>> >>> fe80::200:ff:fe00:0/64 Scope:Link
>>> >>> > >>> >>>           UP BROADCAST
RUNNING MULTICAST
>>> MTU:1500  Metric:1
>>> >>> > >>> >>>           RX packets:43
errors:0 dropped:0
>>> >>> overruns:0 frame:0
>>> >>> > >>> >>>           TX packets:17
errors:0 dropped:0
>>> >>> overruns:0 carrier:0
>>> >>> > >>> >>>           collisions:0
txqueuelen:0
>>> >>> > >>> >>>           RX bytes:3208
(3.1 KiB)  TX
>>> bytes:2018 (1.9 KiB)
>>> >>> > >>> >>>
>>> >>> > >>> >>> I don''t see
any xen bridge, because that''s what I
>>> >>> think I need: one
>>> >>> > >>> >>> network card, and one
xen bridge.
>>> >>> > >>> >>> I found on google that
I could use the following script:
>>> >>> > >>> >>> #!/bin/sh
>>> >>> > >>> >>> dir=$(dirname
"$0")
>>> >>> > >>> >>>
"$dir/network-bridge" "$@" vifnum=0 netdev=eth0
>>> >>> bridge=xenbr0
>>> >>> > >>> >>>
"$dir/network-bridge" "$@" vifnum=1 netdev=eth1
>>> >>> bridge=xenbr1
>>> >>> > >>> >>>
"$dir/network-bridge" "$@" vifnum=2 netdev=eth2
>>> >>> bridge=xenbr2
>>> >>> > >>> >>> And then set it into
xen-config.sxp:
>>> >>> > >>> >>> network-script
matrix-network
>>> >>> > >>> >>> But it gives errors
that network-script has only
>>> >>> start, stop and
>>> >>> > >>> >>> status. The only thing
that it succeds is that I
>>> >>> can see a xen
>>> >>> > >>>
>>> >>> > >>> bridge.
>>> >>> > >>>
>>> >>> > >>> >>> If this would work,
doesn''t this affects other
>>> >>> guest domains also?
>>> >>> > >>> >>>
>>> >>> > >>> >>> My question is: How can
I set a guest dom to use
>>> >>> directly a network
>>> >>> > >>> >>> card with other ip
class ?
>>> >>> > >>> >>>
>>> >>> > >>> >>> Best regards,
>>> >>> > >>> >>> Octav
>>> >>> > >>> >>>
>>> >>> > >>> >>>
>>> >>> > >>> >>>
>>> >>> > >>> >>>
_______________________________________________
>>> >>> > >>> >>> Xen-users mailing list
>>> >>> > >>> >>>
Xen-users@lists.xensource.com
>>> >>> > >>> >>>
http://lists.xensource.com/xen-users
>>> >>> > >>> >>
>>> >>> > >>> >> --
>>> >>> > >>> >> Kobalt W.I.T.
>>> >>> > >>> >> Web & Information
Technology
>>> >>> > >>> >> Brusselsesteenweg 152
>>> >>> > >>> >> 1850 Grimbergen
>>> >>> > >>> >>
>>> >>> > >>> >> Tel  : +32 479 339 655
>>> >>> > >>> >> Email: info@kobaltwit.be
>>> >>> > >>> >>
>>> >>> > >>> >>
_______________________________________________
>>> >>> > >>> >> Xen-users mailing list
>>> >>> > >>> >>
Xen-users@lists.xensource.com
>>> >>> > >>> >>
http://lists.xensource.com/xen-users
>>> >>> > >>
>>> >>> > >> --
>>> >>> > >> Kobalt W.I.T.
>>> >>> > >> Web & Information Technology
>>> >>> > >> Brusselsesteenweg 152
>>> >>> > >> 1850 Grimbergen
>>> >>> > >>
>>> >>> > >> Tel  : +32 479 339 655
>>> >>> > >> Email: info@kobaltwit.be
>>> >>> > >>
>>> >>> > >>
_______________________________________________
>>> >>> > >> Xen-users mailing list
>>> >>> > >> Xen-users@lists.xensource.com
>>> >>> > >> http://lists.xensource.com/xen-users
>>> >>> >
>>> >>> > _______________________________________________
>>> >>> > Xen-users mailing list
>>> >>> > Xen-users@lists.xensource.com
>>> >>> > http://lists.xensource.com/xen-users
>>> >>>
>>> >>> --
>>> >>> Kobalt W.I.T.
>>> >>> Web & Information Technology
>>> >>> Brusselsesteenweg 152
>>> >>> 1850 Grimbergen
>>> >>>
>>> >>> Tel  : +32 479 339 655
>>> >>> Email: info@kobaltwit.be
>>> >>>
>>> >>> _______________________________________________
>>> >>> Xen-users mailing list
>>> >>> Xen-users@lists.xensource.com
>>> >>> http://lists.xensource.com/xen-users
>>> >>>
>>> >>>
>>> >>>
>>> >>
>>> >>
>>> >>
>>> >> _______________________________________________
>>> >> Xen-users mailing list
>>> >> Xen-users@lists.xensource.com
>>> >> http://lists.xensource.com/xen-users
>>> >>
>>>
>>>
>>>
>>>
>>>
>>
>>
>>


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

Nico Kadel-Garcia

2007-Jun-05 20:54 UTC

head link

Re: RHEL 5 glibc, was Re: [Xen-users] Xen a couple of questions

Michael Paesold wrote:> Nico Kadel-Garcia wrote:
>> Geert Janssens wrote:
>>> 1. On CentOS 5, Redhat Enterprise 5 and Fedora core 6 and up, the 
>>> xen kernel can be used for both dom0 and domU. There is no need 
>>> anymore for two kernels.
>>>   
>> Has anyone properly tested the RHEL and CentOS 5 glibc in clients? 
>> I''m forced to use the xensource 3.1.0 kernel-xen kernel to get
RHEL 4
>> DomU''s, and it causes the guests to whine bitterly at boot
time about
>> the /lib/tls issues. Xensource used to publish a modified glibc for 
>> RHEL 4.4 to avoid the /lib/tls emulation issues, but CentOS and RHEL 
>> 4.5 have an even more recent glibc.
>
> What kernel do you use in the guests? The RedHat/CentOS provided 
> kernels properly place glibc hint files into /etc/ld.so.conf.d/, which 
> makes glibc use the nosegneg variants available in RHEL/CentOS 4.5 and 5.
>
> If you use Xensource kernels, I guess those don''t create the files
> correctly.Apparently not! I''ll try some domains with such a kernelcap file and
see
how they do.

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

Mark Williamson

2007-Jun-06 02:48 UTC

head link

Re: [Xen-users] Xen a couple of questions

> 1. Regarding Centos and Fedora core 7 compared with fedora core 5.
I''ve
> seen that on fedora core 5 when you want to install xen you have to
> install the following packages: xen, kernel-xen0 and kernel-xenU (of
> course with the dependencies needed). But on Centos, FC7 and I think
> redhat versions, you only have to install xen and kernel-xen, you
don''t
> have any kernel for the guest system. In my case I could only start a xen
> guest (on FC7) with an older kernel-xenU installed from FC version 5.
>
> My question is: Why does the newer releases of linux has xen kernel
> prebuilt but just for dom0, not for the guest systems, and you
can''t even
> find a domU kernel special for those systems?
The same kernel will work for both; there''s no need to have a different
kernel
for the domUs.

Cheers,
Mark
> 2.My network topology in my home is like this:
> --------
> -router-
> --------
>
>
>
> -----------         ------------
> -linux xen-   ----  -other 2 pc-
> -----------         ------------
>
> The linux xen machine has two network interfaces and xen installed.
> I want: -  one windows machine virtualized
>         -  one linux machine for which I want to have a public ipaddress
> (to put the ip in DMS on the router) and I want it to use eth0 (so
> in this case the traffic can not be sniffed by other guest systems
> or dom0).
>
> ifconfig -a (on dom0) it shows like this:
> eth0      Link encap:Ethernet  HWaddr 00:00:E8:76:E2:4D
>           UP BROADCAST MULTICAST  MTU:1500  Metric:1
>           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:1000
>           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
>           Interrupt:21 Base address:0x2000
>
> eth1      Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>           inet addr:192.168.0.101  Bcast:192.168.0.255  Mask:255.255.255.0
>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:198578 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:117290 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:267328989 (254.9 MiB)  TX bytes:8294632 (7.9 MiB)
>
> lo        Link encap:Local Loopback
>           inet addr:127.0.0.1  Mask:255.0.0.0
>           inet6 addr: ::1/128 Scope:Host
>           UP LOOPBACK RUNNING  MTU:16436  Metric:1
>           RX packets:2689 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:2689 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:12510296 (11.9 MiB)  TX bytes:12510296 (11.9 MiB)
>
> peth1     Link encap:Ethernet  HWaddr 00:16:76:B3:16:AB
>           inet6 addr: fe80::216:76ff:feb3:16ab/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:198588 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:117311 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:100
>           RX bytes:270906777 (258.3 MiB)  TX bytes:8813848 (8.4 MiB)
>           Base address:0x40c0 Memory:92200000-92220000
>
> vif4.0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>           inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:9 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:1 errors:0 dropped:6 overruns:0 carrier:0
>           collisions:0 txqueuelen:1
>           RX bytes:1068 (1.0 KiB)  TX bytes:342 (342.0 b)
>
> virbr0    Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
>           inet addr:192.168.122.1  Bcast:192.168.122.255 
> Mask:255.255.255.0 inet6 addr: fe80::200:ff:fe00:0/64 Scope:Link
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:43 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:17 errors:0 dropped:0 overruns:0 carrier:0
>           collisions:0 txqueuelen:0
>           RX bytes:3208 (3.1 KiB)  TX bytes:2018 (1.9 KiB)
>
> I don''t see any xen bridge, because that''s what I think I
need: one
> network card, and one xen bridge.
> I found on google that I could use the following script:
> #!/bin/sh
> dir=$(dirname "$0")
> "$dir/network-bridge" "$@" vifnum=0 netdev=eth0
bridge=xenbr0
> "$dir/network-bridge" "$@" vifnum=1 netdev=eth1
bridge=xenbr1
> "$dir/network-bridge" "$@" vifnum=2 netdev=eth2
bridge=xenbr2
> And then set it into xen-config.sxp:
> network-script matrix-network
> But it gives errors that network-script has only start, stop and status.
> The only thing that it succeds is that I can see a xen bridge. If this
> would work, doesn''t this affects other guest domains also?
>
> My question is: How can I set a guest dom to use directly a network card
> with other ip class ?
>
> Best regards,
> Octav
>
>
>
> _______________________________________________
> Xen-users mailing list
> Xen-users@lists.xensource.com
> http://lists.xensource.com/xen-users


-- 
Dave: Just a question. What use is a unicyle with no seat?  And no pedals!
Mark: To answer a question with a question: What use is a skateboard?
Dave: Skateboards have wheels.
Mark: My wheel has a wheel!

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

Nico Kadel-Garcia

2007-Jun-06 08:50 UTC

head link

Re: [Xen-users] Xen a couple of questions

Mark Williamson wrote:>> 1. Regarding Centos and Fedora core 7 compared with fedora core 5.
I''ve
>> seen that on fedora core 5 when you want to install xen you have to
>> install the following packages: xen, kernel-xen0 and kernel-xenU (of
>> course with the dependencies needed). But on Centos, FC7 and I think
>> redhat versions, you only have to install xen and kernel-xen, you
don''t
>> have any kernel for the guest system. In my case I could only start a
xen
>> guest (on FC7) with an older kernel-xenU installed from FC version 5.
>>
>> My question is: Why does the newer releases of linux has xen kernel
>> prebuilt but just for dom0, not for the guest systems, and you
can''t even
>> find a domU kernel special for those systems?
>>     
>
> The same kernel will work for both; there''s no need to have a
different kernel
> for the domUs.
>   Hmmph. the same kernel *can* work for both. There may be subtlely 
different behaviors that are beneficial to DomU, although I''m not sure 
in detail what they are: I''ve just run into that sort of thing behavior
in heterogeneous deployments.

No, I think the issue is that CentOS is pegged to RedHat''s kernel 
release model, where an RHEL deployment is supposed to be stable and 
consistent throughout the lifespan of the operating system.  For 
reliable behavior in such an environment, your DomU *must* have a kernel 
as similar as posible to that deployed by RedHat. Dom0 can be forced to 
be more recent to get critical features (shoving Xen Dom0 into a 2.6.9 
kernel is just asking for pain, though.) So Dom0 pretty much needed a 
much newer kernel.

Notice that for RHEL and CentOS 4.5, which now can gracefully be 
installed as DomU''s on top of a 5.0 Dom0, they only have kernel-xenU 
packages, not kernel-xen packages. If you want a 4.5 machine as a Dom0, 
you need to use the xensource kernel or roll your own. And do *not* try 
to backport virt-manager to CentOS 4.5 without being prepared for a lot 
of pain.

There''s also the issue of kernel size: when you''re doing 
micro-deployments (stripped down DomU''s for firewall or similar mini 
setups) there are some advantages to teeny-tiny kernels, and since you 
have a consistent environment of necessary hardware drivers, you can 
actually do it. But it''s a pain to support, and it also lets anyone 
doing a "uname -a" find out that you''re in a Xen guest
environment.

So there are tradeoffs.

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

Petersson, Mats

2007-Jun-06 09:04 UTC

head link

RE: [Xen-users] Xen a couple of questions

> No, I think the issue is that CentOS is pegged to RedHat''s kernel 
> release model, where an RHEL deployment is supposed to be stable and 
> consistent throughout the lifespan of the operating system.  For 
> reliable behavior in such an environment, your DomU *must* 
> have a kernel 
> as similar as posible to that deployed by RedHat. Dom0 can be 
> forced to 
> be more recent to get critical features (shoving Xen Dom0 
> into a 2.6.9 
> kernel is just asking for pain, though.) So Dom0 pretty much needed a 
> much newer kernel.
The default kernel build in current Xen releases (since 3.0.4 or before)
is to build one kernel that works for both. Yes, there are some
differences between a XenU and Xen0 config, in that one has the
"privileged" option set, and there are some differences in drivers
added
to the kernel (particularly, XenU doesn''t normally have ANY support for
ANY hardware drivers). But you should ALWAYS be able to sue the Dom0
kernel for DomU, even if the other way around doesn''t necessarily hold
true. 

> 
> Notice that for RHEL and CentOS 4.5, which now can gracefully be 
> installed as DomU''s on top of a 5.0 Dom0, they only have
kernel-xenU
> packages, not kernel-xen packages. If you want a 4.5 machine 
> as a Dom0, 
> you need to use the xensource kernel or roll your own. And do 
> *not* try 
> to backport virt-manager to CentOS 4.5 without being prepared 
> for a lot 
> of pain.
> 
> There''s also the issue of kernel size: when you''re doing 
> micro-deployments (stripped down DomU''s for firewall or similar
mini
> setups) there are some advantages to teeny-tiny kernels, and 
> since you 
> have a consistent environment of necessary hardware drivers, you can 
> actually do it. But it''s a pain to support, and it also lets
anyone
> doing a "uname -a" find out that you''re in a Xen guest
environment.
Yes, if you are loading dozens or more of guest kernels, the size of the
actual kernel will matter. I''m not sure how much the difference is
tho''.


--
Mats> 
> So there are tradeoffs.
> 
> _______________________________________________
> Xen-users mailing list
> Xen-users@lists.xensource.com
> http://lists.xensource.com/xen-users
> 
> 
> 


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

Nico Kadel-Garcia

2007-Jun-06 09:16 UTC

head link

Re: [Xen-users] Xen a couple of questions

Petersson, Mats wrote:>  
>
>   
>> No, I think the issue is that CentOS is pegged to RedHat''s
kernel
>> release model, where an RHEL deployment is supposed to be stable and 
>> consistent throughout the lifespan of the operating system.  For 
>> reliable behavior in such an environment, your DomU *must* 
>> have a kernel 
>> as similar as posible to that deployed by RedHat. Dom0 can be 
>> forced to 
>> be more recent to get critical features (shoving Xen Dom0 
>> into a 2.6.9 
>> kernel is just asking for pain, though.) So Dom0 pretty much needed a 
>> much newer kernel.
>>     
>
> The default kernel build in current Xen releases (since 3.0.4 or before)
> is to build one kernel that works for both. Yes, there are some
> differences between a XenU and Xen0 config, in that one has the
> "privileged" option set, and there are some differences in
drivers added
> to the kernel (particularly, XenU doesn''t normally have ANY
support for
> ANY hardware drivers). But you should ALWAYS be able to sue the Dom0
> kernel for DomU, even if the other way around doesn''t necessarily
hold
> true. 
>
>
>   
>> Notice that for RHEL and CentOS 4.5, which now can gracefully be 
>> installed as DomU''s on top of a 5.0 Dom0, they only have
kernel-xenU
>> packages, not kernel-xen packages. If you want a 4.5 machine 
>> as a Dom0, 
>> you need to use the xensource kernel or roll your own. And do 
>> *not* try 
>> to backport virt-manager to CentOS 4.5 without being prepared 
>> for a lot 
>> of pain.
>>
>> There''s also the issue of kernel size: when you''re
doing
>> micro-deployments (stripped down DomU''s for firewall or
similar mini
>> setups) there are some advantages to teeny-tiny kernels, and 
>> since you 
>> have a consistent environment of necessary hardware drivers, you can 
>> actually do it. But it''s a pain to support, and it also lets
anyone
>> doing a "uname -a" find out that you''re in a Xen
guest environment.
>>     
>
> Yes, if you are loading dozens or more of guest kernels, the size of the
> actual kernel will matter. I''m not sure how much the difference is
tho''.
>   And if you call RedHat, Oracle, or McAfee for support with kernel 
related issues on RHEL 4, and tell them "I''m running a 2.6.18
kernel",
they''re going to have good reason to throw p their hands and say
"revert
to the published kernel, then we can help you". They might not: premium 
corporate support is pretty good, but I''d be tempted to do that as the 
support person.

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

Petersson, Mats

2007-Jun-06 09:33 UTC

head link

RE: [Xen-users] Xen a couple of questions

> >
> > Yes, if you are loading dozens or more of guest kernels, 
> the size of the
> > actual kernel will matter. I''m not sure how much the 
> difference is tho''.
> >   
> And if you call RedHat, Oracle, or McAfee for support with kernel 
> related issues on RHEL 4, and tell them "I''m running a 2.6.18
> kernel", 
> they''re going to have good reason to throw p their hands and 
> say "revert 
> to the published kernel, then we can help you". They might 
> not: premium 
> corporate support is pretty good, but I''d be tempted to do 
> that as the 
> support person.
Hmm, and you don''t think a 2.6.9-XenU kernel will have the same
"not
original kernel" status?

--
Mats



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

Nico Kadel-Garcia

2007-Jun-06 11:07 UTC

head link

Re: [Xen-users] Xen a couple of questions

Petersson, Mats wrote:>>> Yes, if you are loading dozens or more of guest kernels, 
>>>       
>> the size of the
>>     
>>> actual kernel will matter. I''m not sure how much the 
>>>       
>> difference is tho''.
>>     
>>>   
>>>       
>> And if you call RedHat, Oracle, or McAfee for support with kernel 
>> related issues on RHEL 4, and tell them "I''m running a
2.6.18
>> kernel", 
>> they''re going to have good reason to throw p their hands and 
>> say "revert 
>> to the published kernel, then we can help you". They might 
>> not: premium 
>> corporate support is pretty good, but I''d be tempted to do 
>> that as the 
>> support person.
>>     
>
> Hmm, and you don''t think a 2.6.9-XenU kernel will have the same
"not
> original kernel" status?
>   It''s now available as part of the base RHEL 4.5 installation, which was
released less than 2 months ago. It''s superior to the xensource kernel 
in some ways, since it interacts wiell with the updated grub and allows 
the graceful use of pygrub without having to hand-edit your grub.conf 
files before rebooting, and it drops a "kernelcap" file in 
/etc/ld.so.conf.d/ to avoid the TLS emulation problems. Plus it has the 
documentation with it, which the xensource RPM does not and for which 
building from SRPM fails.

So it''s a fairly good kernel. Too bad the network fails between Dom0
and
DomU with it, which means I''m still sticking with the Xensource 3.1.0 
kernels for my guests.

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

Mark Williamson

2007-Jun-06 19:42 UTC

head link

Re: [Xen-users] Xen a couple of questions

> >> 1. Regarding Centos and Fedora core 7 compared with fedora core 5.
I''ve
> >> seen that on fedora core 5 when you want to install xen you have
to
> >> install the following packages: xen, kernel-xen0 and kernel-xenU
(of
> >> course with the dependencies needed). But on Centos, FC7 and I
think
> >> redhat versions, you only have to install xen and kernel-xen, you
don''t
> >> have any kernel for the guest system. In my case I could only
start a
> >> xen guest (on FC7) with an older kernel-xenU installed from FC
version
> >> 5.
> >>
> >> My question is: Why does the newer releases of linux has xen
kernel
> >> prebuilt but just for dom0, not for the guest systems, and you
can''t
> >> even find a domU kernel special for those systems?
> >
> > The same kernel will work for both; there''s no need to have a
different
> > kernel for the domUs.
>
> Hmmph. the same kernel *can* work for both. There may be subtlely
> different behaviors that are beneficial to DomU, although I''m not
sure
> in detail what they are: I''ve just run into that sort of thing
behavior
> in heterogeneous deployments.
>
> No, I think the issue is that CentOS is pegged to RedHat''s kernel
> release model, where an RHEL deployment is supposed to be stable and
> consistent throughout the lifespan of the operating system.  For
> reliable behavior in such an environment, your DomU *must* have a kernel
> as similar as posible to that deployed by RedHat. Dom0 can be forced to
> be more recent to get critical features (shoving Xen Dom0 into a 2.6.9
> kernel is just asking for pain, though.) So Dom0 pretty much needed a
> much newer kernel.
Sure.  I was referring to the specific case of Fedora 7 and why there
weren''t
separate kernel packages for dom0 and domU - you should use the same kernel 
for both (in the case that you''re running a Fedora guest).

If you''re running a range of different distros, the best plan is to run
each
install on that distro''s Xenified kernel rather than trying to use the
same
kernel for each of them.
> Notice that for RHEL and CentOS 4.5, which now can gracefully be
> installed as DomU''s on top of a 5.0 Dom0, they only have
kernel-xenU
> packages, not kernel-xen packages. If you want a 4.5 machine as a Dom0,
> you need to use the xensource kernel or roll your own. And do *not* try
> to backport virt-manager to CentOS 4.5 without being prepared for a lot
> of pain.
Been there, done that :-(

I just use commandline tools on CentOS 4.x, and have built my own kernels for 
all the domains.  But that''s my development box, so this is what
I''d do
anyhow.
> There''s also the issue of kernel size: when you''re doing
> micro-deployments (stripped down DomU''s for firewall or similar
mini
> setups) there are some advantages to teeny-tiny kernels, and since you
> have a consistent environment of necessary hardware drivers, you can
> actually do it. But it''s a pain to support, and it also lets
anyone
> doing a "uname -a" find out that you''re in a Xen guest
environment.
>
> So there are tradeoffs.
*nod*

But if a distro supports a single -xen kernel rather than -xen0 and -xenU then 
it''s nothing to worry about; the extra cruft required by dom0 is mostly
device drivers as modules, so you don''t actually need them in memory.

The nicest way to handle this is probably to boot all the guests using pygrub 
and let them decide which kernel they want to boot off their own filesystem.  
The guest OS can be responsible for updating this kernel as necessary, etc.

Cheers,
Mark

-- 
Dave: Just a question. What use is a unicyle with no seat?  And no pedals!
Mark: To answer a question with a question: What use is a skateboard?
Dave: Skateboards have wheels.
Mark: My wheel has a wheel!

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

Xen users - Jun 2007 - Xen a couple of questions

[Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

RHEL 5 glibc, was Re: [Xen-users] Xen a couple of questions

Re: RHEL 5 glibc, was Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

RE: [Xen-users] Xen a couple of questions

RE: [Xen-users] Xen a couple of questions

RE: [Xen-users] Xen a couple of questions

RE: [Xen-users] Xen a couple of questions

RE: [Xen-users] Xen a couple of questions

Re: RHEL 5 glibc, was Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

RE: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

RE: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions

Re: [Xen-users] Xen a couple of questions