[ipxe-devel] Official ipxe.efi fails to chain the 1st time into boot.salstar.sk

[Christian Nilsson]

tcpdump -i br0 src 192.168.1.16 -w tcpdump-w.test2.pcap
Only captures packets from 192.168.1.16, not the packets that comes in
return.
instead try
tcpdump -i br0 host 192.168.1.16 -w tcpdump-w.test2.pcap

the error url http://ipxe.org/err/4c0a60 tells us there is a connection
issue, possibly routing or firewall.

Would be interesting to know if
chain http://boot.salstar.sk/ipxe/ipxe.efi || chain
http://boot.salstar.sk/ipxe/ipxe.efi
works, it will do the same thing but directly after each other.
You might also want to try and see if you can recreate this issue with a
browser on the same network. And also doing a tracert/traceroute at the
same time to boot.salstar.sk to see if it gives differing results between
working and non working. (or if it differs when being run several times
after each other)

On Sat, Sep 24, 2016 at 10:11 PM, Dreamcat4 <dreamcat4 at gmail.com> wrote:

> The cmd used to create the pcap file was (on my LAN router):
>
> tcpdump -i br0 src 92.168.1.16 -w tcpdump-w.test2.pcap
>
>
> This is what happened in the ipxe shell:
>
> Type 'exit' to get back to the menu
> iPXE> chain http://boot.salstar.sk/ipxe/ipxe.efi
> http://boot.salstar.sk/ipxe/ipxe.efi........ Connection timed out (
> http://ipxe.org/4c0a6092)
> iPXE> _
>
> # Again (up arrow) 2nd time successful
>
> iPXE> chain http://boot.salstar.sk/ipxe/ipxe.efi... ok
> iPXE initialising devices...
>
> ========
> end of pcap logging.
>
>
> Sounds like this problem:
>
> http://lists.ipxe.org/pipermail/ipxe-devel/2014-July/003627.html
>
> But I dont have any special Cisco switch or other special hardware. So i
> reboot and try again to a different web server. This time ipxe.org's
> website and not salstar.sk.
>
> This time it worked 1st time:
>
> Type 'exit' to get back to the menu
> iPXE> chain http://boot.ipxe.org/ipxe.efi... ok
> iPXE initialising devices...
>
> Which worked the 1st time no problems. No connection timeout whatsoever.
>
> Earlier today I asked Jan, the site maintainer of salstar.sk to help find
> the issue. He took a packet log his end and didnt see anything failed come
> through apart from the successful 200 ok one.
>
> This was during a different run.
>
> [Copied from his email]
>
> One try only:
>
> GET /ipxe/ipxe.efi HTTP/1.1
>
> Host: boot.salstar.sk
>
> User-Agent: iPXE/1.0.0+ (827dd)
>
> Connection: keep-alive
>
>
>
> HTTP/1.1 200 OK
>
> Date: Sat, 24 Sep 2016 18:22:15 GMT
>
> Server: Apache/2.4.23 (Fedora) OpenSSL/1.0.2h-fips PHP/5.6.25
>
> Upgrade: h2,h2c
>
> Connection: Upgrade, close
>
> Last-Modified: Sat, 24 Sep 2016 06:23:22 GMT
>
> ETag: "e9ce0-53d3aee7d0579"
>
> Accept-Ranges: bytes
>
> Content-Length: 957664
>
> Cache-Control: max-age=600, private
>
> Content-Type: application/efi
>
>
> I think there is problem on your side.
> Try to sniff some packets at your side, there are no packets from you
> here.
>
>
>
> On Sat, Sep 24, 2016 at 9:04 PM, Christian Nilsson <nikize at gmail.com>
> wrote:
>
>> That "blah error url" is one of the most important pieces of information.
>> The pcap file only has the part where 192.168.1.16 sends packets, and
>> nothing is received. Did you filter out a bit to much from the pcap?
>>
>> Please send that error url as a response to your earlier email to the
>> mailing list and we can start taking a look.
>> Also I tried to respond on IRC but not sure if you got that message?
>>
>> On Sat, Sep 24, 2016 at 9:37 PM, Dreamcat4 <dreamcat4 at gmail.com> wrote:
>>
>>> Hi,
>>>
>>> We tried this in shell:
>>>
>>> iPXE> chain http://boot.salstar.sk/ipxe/ipxe.efi
>>> http://boot.salstar.sk/ipxe/ipxe.efi.... connection timed out (blah
>>> error url)
>>>
>>> iPXE> chain http://boot.salstar.sk/ipxe/ipxe.efi
>>> http://boot.salstar.sk/ipxe/ipxe.efi.... ok
>>>
>>> 2nd time worked.
>>>
>>> Nothing is seen on the remote webserver the 1st failed attempt. And
>>> other regular local PC client (eg wget on linux) can download the same file
>>> no problems.
>>>
>>> Other urls worked fine the 1st time. For example
>>>
>>> iPXE> chain http://boot.ipxe.org/ipxe.efi
>>>
>>> I attached to this email a .pcap file of the local packet trace log file.
>>>
>>> Its a problem because the 1st connection timeout error (or any other
>>> error) and my ipxe boot menu will exit / abort. It never (from menu script)
>>> gets the change to try again for a 2nd time.
>>>
>>>
>>> Other times it will fail up to 4 times, then work ok. But in the pcap
>>> log file attached here, it worked the 2nd time.
>>>
>>>
>>>
>>> _______________________________________________
>>> ipxe-devel mailing list
>>> ipxe-devel at lists.ipxe.org
>>> https://lists.ipxe.org/mailman/listinfo.cgi/ipxe-devel
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ipxe.org/pipermail/ipxe-devel/attachments/20160924/2b00f81b/attachment.htm>