[Users] ovirt 3.1 guest pxe boot fails

[Rami Vaknin]

On 10/09/2012 01:23 PM, Yuriy Demchenko wrote:
> Hi,
>
> I've checked this issue with ipxe-roms-qemu from default fedora repo 
> (ipxe-roms-qemu-20120328-1.gitaac9718.fc17) - same behaviour: first 
> time system pxe boots fine after domain start, but after guest-reboot 
> - pxe boot fails, cant get ip from dhcp and malformed packets on 
> bridge interface.
> And same situation with fedora host not connected to ovirt (vm created 
> with virt-manager)

What is the nic type (virtio, e1000, ...)? could you please try with 
another type too?
>
> Yuriy Demchenko
>
> On 10/05/2012 11:55 AM, Yuriy Demchenko wrote:
>> Hi,
>>
>> I cant use ipxe-roms-qemu-20120328-1.gitaac9718.fc17 from default 
>> repo (cant boot from SAN with it - http://ipxe.org/err/1d704439), so 
>> i've took srpm and rebuild it with lastest git sources
>> so now i'm using ipxe-roms-qemu-20120926-1.git09cc63f.fc17
>>
>> still cant find logical connection between ipxe version and my 
>> problem, as first boot never fails, only reboots initiated from guest 
>> system does; same goes for real hardware - i chainload same ipxe rom 
>> into real machine and it boots/reboots fine. dunno about qemu/kvm not 
>> connected to ovirt - i'll check it out on a weekend.
>>
>> Yuriy Demchenko
>>
>> On 10/05/2012 01:16 AM, Rami Vaknin wrote:
>>> On 10/04/2012 12:49 PM, Yuriy Demchenko wrote:
>>>> Hi,
>>>>
>>>> I've started testing ovirt 3.1 in my lab environment and found some 
>>>> strange behaviour/bug:
>>>> when i create vm and set it to boot from network - first time it 
>>>> boots fine
>>>>> 14:23:43.718613 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, 
>>>>> Request from 52:54:00:c7:76:50, length 395
>>>
>>> Although you managed to boot from network in the first time, I'm not 
>>> sure that DHCP packet with length of 395 is ok - I think it should 
>>> be ~60, I would start with checking the gpxe-roms-qemu package version.
>>>
>>>> pxe works, ip address received and machine boots,
>>>> but if I'll restart it via guest command 'reboot' (i.e not through 
>>>> stop/start from ovirt admin interface) - pxe boot fails, ipxe just 
>>>> states timeout on dhcp requests
>>>> and with tcpdump i even cannot see any dhcp requests at all!
>>>> instead of that i get some weird malformed packets:
>>>>> 14:21:38.529244 00:00:00:00:ff:ff > 00:00:00:00:00:00, ethertype 
>>>>> Unknown (0xffff), length 447:
>>>>>     0x0000:  ffff 5254 00c7 7650 0800 4500 01a7 0322 ..RT..vP..E...."
>>>>>     0x0010:  0000 4011 7625 0000 0000 ffff ffff 0044 .. at .v%.........D
>>>>>     0x0020:  0043 0193 02f7 0101 0600 2c78 5053 000c .C........,xPS..
>>>>>     0x0030:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x0040:  0000 5254 00c7 7650 0000 0000 0000 0000 ..RT..vP........
>>>>>     0x0050:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x0060:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x0070:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x0080:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x0090:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x00a0:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x00b0:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x00c0:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x00d0:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x00e0:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x00f0:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x0100:  0000 0000 0000 0000 0000 0000 0000 0000 ................
>>>>>     0x0110:  0000 6382 5363 3501 0139 0205 c05d 0200 ..c.Sc5..9...]..
>>>>>     0x0120:  005e 0301 0201 3c20 5058 4543 6c69 656e .^....<.PXEClien
>>>>>     0x0130:  743a 4172 6368 3a30 3030 3030 3a55 4e44 t:Arch:00000:UND
>>>>>     0x0140:  493a 3030 3230 3031 4d04 6950 5845 3715 I:002001M.iPXE7.
>>>>>     0x0150:  0103 0607 0c0f 112b 3c42 4380 8182 8384 .......+<BC.....
>>>>>     0x0160:  8586 87af cbaf 2db1 0501 1af4 1000 2201 ......-.......".
>>>>>     0x0170:  0119 0101 2101 0118 0101 1101 0113 0101 ....!...........
>>>>>     0x0180:  1701 0115 0101 2701 0110 0102 1201 01eb ......'.........
>>>>>     0x0190:  0301 0000 3d07 0152 5400 c776 5061 1100 ....=..RT..vPa..
>>>>>     0x01a0:  21b3 bc0f e882 4017 9c20 f3e5 0afd 0547 !..... at ........G
>>>>>     0x01b0:  ff                                       .
>>>>
>>>> As i'm planning to heavily rely on network booting this issue is 
>>>> critical for me, any tips how to solve it?
>>>>
>>>> ovirt 3.1, engine and node on separate hosts, installed on clean 
>>>> fedora 17/minimal, node with kernel-3.3.4-5.fc17.x86_64
>>>> node has 1 nic, 1 network defined in ovirt - ovirtmgmt
>>>>
>>>
>>>
>>
>


-- 

Thanks,

Rami Vaknin, QE @ Red Hat, TLV, IL.