On 11/14/2011 07:47 AM, Wizard wrote:
> 2011/11/7 Lucas Meneghel Rodrigues<[email protected]>:
>> On 11/07/2011 10:00 AM, Wizard wrote:
>>>
>>> 2011/11/7 Lucas Meneghel Rodrigues<[email protected]>:
>>>>
>>>> On 11/05/2011 12:54 PM, Wizard wrote:
>>>>>
>>>>> 2011/11/5 Lucas Meneghel Rodrigues<[email protected]>:
>>>>>>
>>>>>> On 11/05/2011 06:57 AM, Wizard wrote:
>>>>>>>
>>>>>>> 2011/11/4 Lucas Meneghel Rodrigues<[email protected]>:
>>>>>>>>
>>>>>>>> On 11/04/2011 12:49 PM, Wizard wrote:
>>>>>>>>>
>>>>>>>>> Lucas
>>>>>>>>>
>>>>>>>>> I read the tips for create a uptime test case for autotest.
>>>>>>>>>
>>>>>>>>> But I faced an error message.
>>>>>>>>>
>>>>>>>>> The config is :
>>>>>>>>> ../../common_lib/cartesian_config.py tests.cfg
>>>>>>>>> dict    1:  smp2.CustomGuestLinux.uptime
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> The command of qemu is :
>>>>>>>>> /usr/bin/qemu -name 'vm1' -nodefaults -vga std -monitor
>>>>>>>>> unix:'/tmp/monitor-humanmonitor1-20111104-223259-kgX2',server,nowait
>>>>>>>>> -serial unix:'/tmp/serial-20111104-223259-kgX2',server,nowait -drive
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> file='/home/richard/kvm/image/custom_image_linux',index=0,if=ide,cache=none
>>>>>>>>> -device
>>>>>>>>> rtl8139,netdev=idLRqDTg,mac='9a:64:5d:40:fb:fa',id='idsYMlt4'
>>>>>>>>> -netdev tap,id=idLRqDTg,fd=22 -m 1024 -smp 2 -vnc :0
>>>>>>>>>
>>>>>>>>> The error is:
>>>>>>>>> MissingError: Cannot find IP address for MAC address
>>>>>>>>> 9a:64:5d:40:fb:fa
>>>>>>>>>     [context: logging into 'vm1']
>>>>>>>>
>>>>>>>> You probably want to connect on your vm's vnc session. This means
>>>>>>>> your
>>>>>>>> linux
>>>>>>>> guest did not even try to get an IP from the DHCP server.
>>>>>>>>
>>>>>>>> vncviewer localhost:0
>>>>>>>>
>>>>>>> I tried vncviewer :0 , and can see the guest starts up and logged in
>>>>>>> successfully.
>>>>>>>
>>>>>>>> Also, kvm autotest produces screenshots, that go into
>>>>>>>> client/results/default/[your test name]screendumps_vm1. As this is a
>>>>>>>> custom
>>>>>>>> image, it's very hard to tell what is going wrong, but usually means
>>>>>>>> the
>>>>>>>> boot got stuck somewhere and your linux guest did not bring up a
>>>>>>>> newtork
>>>>>>>> interface.
>>>>>>>>
>>>>>>>
>>>>>>> But the autotest finally failed with the message above.
>>>>>>> I guess this error is printed when autotest client try to connect
>>>>>>> guest through ssh
>>>>>>> session.
>>>>>>
>>>>>> Well, just to cover all bases - is the guest configured to pick an IP
>>>>>> address from a DHCP server, or it's configured with static IP and DNS?
>>>>>> When
>>>>>> you say logged in, you probably mean "I can see a getty login prompt",
>>>>>> which
>>>>>> doesn't mean much for KVM autotest remote session.
>>>>>>
>>>>> You are right. The guest use static ip address.
>>>>>
>>>>> But, one curious thing is after I change the sysconfig file.
>>>>> I use "qemu rhel.img" to bootup, log in, I can see the ip address of
>>>>> 10.0.2.15.
>>>>> While I use autotest, I use vnc to view the guest, I don't see the
>>>>> interface is up.
>>>>
>>>> Because KVM autotest uses TAP, and your command, qemu rhel.img, makes
>>>> qemu
>>>> use userspace networing. Userspace (also known as slirp) means qemu will
>>>> provide an internal network for the guest, with a built in DHCP server.
>>>> 10.0.2.1 is the DHCP server, and your guest will be assigned a DHCP
>>>> address
>>>> on this range. Therefore, by default with KVM autotest you will not see
>>>> this
>>>> interface up.
>>>>
>>>> Slirp suffers from bugs that can potentially crash qemu, so it's not a
>>>> supported and/or reliable option, and that is why KVM autotest provides
>>>> TAP
>>>> by default.
>>>>
>>> Ok, it use two different network mode. While, how could I make the
>>> guest in the KVM autotest to have an ip address?
>>>
>>> What should I config?
>>>
>>> The libvirt is running and virbr0 is up.
>>> The guest is configured with dhcp enabled.
>>
>> Then you just have to make sure the password is the same as the password KVM
>> autotest expects, which is 123456, as stated on the config file
>> guest-os.cfg. Also, another thing that would be reasonable to expect is that
>> sshd is up and listening on port 22 and firewall allow access to that
>> service.
>>
>> It seems you are using RHEL, so SELinux might play some pranks on you,
>> although it is hard to anticipate all things that could go wrong. but well,
>> if you check the things I stated on the first sentence, I believe it will
>> work.
>>
> Hmm... After the client strat up, I log in the guest and see the sshd
> is running and I stop the iptables.
> Still get the same error message.
>
> I should disable the iptables at boot?

That's a good idea.

> BTW, I log in the guest, but see no ip address for eth0.

So try again with firewall disabled.
_______________________________________________
Autotest mailing list
[email protected]
http://test.kernel.org/cgi-bin/mailman/listinfo/autotest

Reply via email to