tftp - bad checksum error? can't transfer file
Anton Shterenlikht
% grep boot /etc/inetd.conf
tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
bootps dgram udp wait root /usr/libexec/bootpd bootpd -d4 -t0 /etc/bootptab
%
I'm trying to boot another node via bootp/tftp.
I'm monitoring what's going on with tcpdump -vv.
bootpd starts fine, and seems to read /etc/bootptab
correctly.
But after that nothing happens, and after a few minutes
I get on the console of the host which I want to boot:
> bootp()/bsd.rd.IP32
Unable to execute bootp()/bsd.rd.IP32: invalid argument
>
Does this tcpdump fragment help to understand where
the problem is:
192.168.232.10.15388 > buzi.tftp: [no cksum] 25 RRQ "/bsd.rd.IP32" octet (
o
23:25:21.024160 IP (tos 0x0, ttl 64, id 56, offset 0, flags [none], proto UDP (1
7), length 30, bad cksum 0 (->293a)!)
buzi.19330 > 192.168.232.10.15388: [udp sum ok] UDP, length 2
23:25:51.013759 IP (tos 0x0, ttl 255, id 256, offset 0, flags [none], proto UDP
(17), length 53)
I'm happy to provide more detailed information
on the settings I use and a longer tcpdump output,
if this helps.
Many thanks
Anton
--
Anton Shterenlikht
Room 2.6, Queen's Building
Mech Eng Dept
Bristol University
University Walk, Bristol BS8 1TR, UK
Tel: +44 (0)117 331 5944
Fax: +44 (0)117 929 4423
_______________________________________________
freebsd-...@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "freebsd-questi...@freebsd.org"
Gary Gatten
Check your file names (case), perms, etc. 90% of time I typo a name or forget to chmod the files; or when using tftp to write I forget to create a file with the correct name (and perms) first.
Many thanks
Anton
<font size="1">
<div style='border:none;border-bottom:double windowtext 2.25pt;padding:0in 0in 1.0pt 0in'>
</div>
"This email is intended to be reviewed by only the intended recipient
and may contain information that is privileged and/or confidential.
If you are not the intended recipient, you are hereby notified that
any review, use, dissemination, disclosure or copying of this email
and its attachments, if any, is strictly prohibited. If you have
received this email in error, please immediately notify the sender by
return email and delete this email from your system."
</font>
Anton Shterenlikht
On Thu, Jul 07, 2011 at 05:54:38PM -0500, Gary Gatten wrote:
> >From my experience tcpdump is misleading re udp fragments and chksums.
> If the packet gets fragmented, udp will report bad chksums at some point.
>
> Check your file names (case), perms, etc.
> 90% of time I typo a name or forget to chmod the files;
% ls -al /tftpboot/bsd.rd.IP32
-rwxrwxrwx 1 root wheel 9646005 Jul 7 21:02 /tftpboot/bsd.rd.IP32
%
% cat /etc/bootptab
o2:\
:ht=ether:\
:ha=0800690c0e3f:\
:sm=255.255.255.0:\
:ip=192.168.232.10:\
:td=/tftpboot:bf=bsd.rd.IP32:
%
Does this look ok?
C. P. Ghost
> I'm trying to troubleshoot tftpd(8).
(...)
> 192.168.232.10.15388 > buzi.tftp: [no cksum] 25 RRQ "/bsd.rd.IP32" octet (
> o
> 23:25:21.024160 IP (tos 0x0, ttl 64, id 56, offset 0, flags [none], proto UDP (1
> 7), length 30, bad cksum 0 (->293a)!)
> buzi.19330 > 192.168.232.10.15388: [udp sum ok] UDP, length 2
> 23:25:51.013759 IP (tos 0x0, ttl 255, id 256, offset 0, flags [none], proto UDP
> (17), length 53)
I've had some trouble netbooting / jumpstaring recently with a similar
pattern (using RARP/BOOTP/TFTP/NFS). It turned out to be a dying
port on the switch whose errors were masked by TCP in day to day
use, but alas were too frequent for UDP.
-cpghost.
--
Cordula's Web. http://www.cordula.ws/
Anton Shterenlikht
> On Fri, Jul 8, 2011 at 12:44 AM, Anton Shterenlikht <me...@bristol.ac.uk> wrote:
> > I'm trying to troubleshoot tftpd(8).
>
> (...)
>
> > o
> > 23:25:21.024160 IP (tos 0x0, ttl 64, id 56, offset 0, flags [none], proto UDP (1
> > 7), length 30, bad cksum 0 (->293a)!)
> > 23:25:51.013759 IP (tos 0x0, ttl 255, id 256, offset 0, flags [none], proto UDP
> > (17), length 53)
>
> I've had some trouble netbooting / jumpstaring recently with a similar
> pattern (using RARP/BOOTP/TFTP/NFS). It turned out to be a dying
> port on the switch whose errors were masked by TCP in day to day
> use, but alas were too frequent for UDP.
I just have a direct ethernet connection
between my FreeBSD laptop with bootpd/tftpd
servers (one ethernet port only) and a node
which I want to boot (also a single ethernet
port). Is there way for me to check whether
either of these ports are "dying"?
Any further diagnostics I can do?
Many thanks
Anton
--
Anton Shterenlikht
Room 2.6, Queen's Building
Mech Eng Dept
Bristol University
University Walk, Bristol BS8 1TR, UK
Tel: +44 (0)117 331 5944
Fax: +44 (0)117 929 4423
Warren Block
> I'm trying to troubleshoot tftpd(8).
>
> % grep boot /etc/inetd.conf
> tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
> tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
> bootps dgram udp wait root /usr/libexec/bootpd bootpd -d4 -t0 /etc/bootptab
> %
>
> I'm trying to boot another node via bootp/tftp.
>
> I'm monitoring what's going on with tcpdump -vv.
>
> bootpd starts fine, and seems to read /etc/bootptab
> correctly.
>
> But after that nothing happens, and after a few minutes
> I get on the console of the host which I want to boot:
>
>> bootp()/bsd.rd.IP32
> Unable to execute bootp()/bsd.rd.IP32: invalid argument
How big is bsd.rd.IP32? Does the BUGS section of tftpd(8) apply?
If so, try ftp/tftp-hpa. It's faster, too, AFAIR.
C. P. Ghost
>> I've had some trouble netbooting / jumpstaring recently with a similar
>> pattern (using RARP/BOOTP/TFTP/NFS). It turned out to be a dying
>> port on the switch whose errors were masked by TCP in day to day
>> use, but alas were too frequent for UDP.
>
> I just have a direct ethernet connection
> between my FreeBSD laptop with bootpd/tftpd
> servers (one ethernet port only) and a node
> which I want to boot (also a single ethernet
> port). Is there way for me to check whether
> either of these ports are "dying"?
> Any further diagnostics I can do?
Just push a lot of data (/dev/zero, /dev/urandom, ...) over
this connection, via ssh or something like that, and monitor
the error rates in "netstat -in" on both ends. Look at Ierrs,
Idrop and Oerrs columns. If one of those ports are dying,
or if your cable isn't properly shielded, you'll notice immediately.
> Many thanks
> Anton
-cpghost.
--
Cordula's Web. http://www.cordula.ws/
Marco Steinbach
> I'm trying to troubleshoot tftpd(8).
>
> % grep boot /etc/inetd.conf
> tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
> tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
> bootps dgram udp wait root /usr/libexec/bootpd bootpd -d4 -t0 /etc/bootptab
> %
>
> I'm trying to boot another node via bootp/tftp.
>
> I'm monitoring what's going on with tcpdump -vv.
>
> bootpd starts fine, and seems to read /etc/bootptab
> correctly.
>
> But after that nothing happens, and after a few minutes
> I get on the console of the host which I want to boot:
>
>> bootp()/bsd.rd.IP32
> Unable to execute bootp()/bsd.rd.IP32: invalid argument
>
Last time I saw this error was with an SGI Octance 2, where the Octanes
RRQ requests contained trailing garbage, which FreeBSDs tftp tried to
interpret as extended options.
Try disabling options_rfc and options_extra in
/usr/src/libexec/tftpd/tftp-options.c, and reinstall tftpd.
MfG CoCo
Marco Steinbach
> Anton Shterenlikht schrieb:
>> I'm trying to troubleshoot tftpd(8).
>>
>> % grep boot /etc/inetd.conf
>> tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
>> tftp dgram udp wait root /usr/libexec/tftpd tftpd -dd -l -s /tftpboot
>> bootps dgram udp wait root /usr/libexec/bootpd bootpd -d4 -t0
>> /etc/bootptab
>> %
>>
>> I'm trying to boot another node via bootp/tftp.
>>
>> I'm monitoring what's going on with tcpdump -vv.
>>
>> bootpd starts fine, and seems to read /etc/bootptab
>> correctly.
>>
>> But after that nothing happens, and after a few minutes
>> I get on the console of the host which I want to boot:
>>
>>> bootp()/bsd.rd.IP32
>> Unable to execute bootp()/bsd.rd.IP32: invalid argument
>>
> [...]
>
> Last time I saw this error was with an SGI Octance 2, where the Octanes
> RRQ requests contained trailing garbage, which FreeBSDs tftp tried to
> interpret as extended options.
>
> Try disabling options_rfc and options_extra in
> /usr/src/libexec/tftpd/tftp-options.c, and reinstall tftpd.
>
Here's another, probably more feasible suggestion:
tftpd(8)s man page (on 8.2-STABLE) tells:
-o Disable support for RFC2347 style TFTP Options.
I don't remember, if that knob didn't do the trick for me, or if I just
didn't see and try it. I assume the latter.