Well, googling has failed me. We've got a pxeboot setup, but we keep getting, in the logs, parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length. This is CentOS 6.2
The section for those machines that can pxeboot is: group { allow booting; allow bootp; filename "pxelinux.0"; option dhcp.bootfile-name "pxelinux.0"; (I added this, based on something I found googling) option subnet-mask 255.255.254.0; option routers <our subnet elided>; default-lease-time 172800; # 2 days. max-lease-time 432000; # 5 days. next-server vessel.cit.nih.gov; #BOOT THESE
I see the entries for option-128 and option-129, but those, based on still more googling, seem to actually be for etherboot, not pxeboot.
So, does anyone have any idea where this buffer length is defined?
mark, one of whose goals is less garbage in the logs
m.roth@5-cent.us writes:
Well, googling has failed me. We've got a pxeboot setup, but we keep getting, in the logs, parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length. This is CentOS 6.2
No such option is listed in the dhcpd.conf man page. Get rid of it, the filename directive before that is all you need.
Lars Hecking wrote:
m.roth@5-cent.us writes:
Well, googling has failed me. We've got a pxeboot setup, but we keep getting, in the logs, parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length. This is CentOS 6.2
No such option is listed in the dhcpd.conf man page. Get rid of it, the filename directive before that is all you need.
That's fine... but getting rid of that won't make the original annoyance go away.
mark
On 04/16/2012 09:04 AM, m.roth@5-cent.us wrote:
Lars Hecking wrote:
m.roth@5-cent.us writes:
Well, googling has failed me. We've got a pxeboot setup, but we keep getting, in the logs, parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length. This is CentOS 6.2
No such option is listed in the dhcpd.conf man page. Get rid of it, the filename directive before that is all you need.
That's fine... but getting rid of that won't make the original annoyance go away.
If you remove the dhcp.bootfile-name option line, how will that not fix your problem?
Johnny Hughes wrote:
On 04/16/2012 09:04 AM, m.roth@5-cent.us wrote:
Lars Hecking wrote:
m.roth@5-cent.us writes:
Well, googling has failed me. We've got a pxeboot setup, but we keep getting, in the logs, parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length. This is CentOS 6.2
No such option is listed in the dhcpd.conf man page. Get rid of it, the filename directive before that is all you need.
That's fine... but getting rid of that won't make the original annoyance go away.
If you remove the dhcp.bootfile-name option line, how will that not fix your problem?
I just added that Friday. We've been getting the original error message for weeks? months? for the configuration *without* option dhcp.bootfile-name
For example, Apr 2 13:45:44 <hostname> dhcpd: parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length.
which is the annoyance I was trying to fix. At that time, the dhcpd.conf section for pxebooting was group { allow booting; allow bootp; filename "pxelinux.0"; option subnet-mask 255.255.254.0; option routers <subnet IP elided>; default-lease-time 172800; # 2 days. max-lease-time 432000; # 5 days.
mark
For example, Apr 2 13:45:44 <hostname> dhcpd: parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length.
Could be a request by a misconfigured client. Check the dhcp client config of clients that are logged at around the same time as the error above.
Lars Hecking wrote:
For example, Apr 2 13:45:44 <hostname> dhcpd: parse_option_buffer: malformed option dhcp.bootfile-name (code 67): option length exceeds option buffer length.
Could be a request by a misconfigured client. Check the dhcp client config of clients that are logged at around the same time as the error
above.
Can't see how - one example of a client is /etc/dhcp/dhclient-eth0.conf: send vendor-class-identifier "anaconda-Linux 2.6.32-71.el6.x86_64 x86_64";
and what's in /etc/dhcp/dhclient.d is -rwxr-xr-x. 1 root root 3221 Nov 11 2010 nis.sh -rwxr-xr-x. 1 root root 1944 May 20 2009 ntp.sh
What kind of configuration were you thinking of that could be wrong?
mark
Can't see how - one example of a client is /etc/dhcp/dhclient-eth0.conf: send vendor-class-identifier "anaconda-Linux 2.6.32-71.el6.x86_64 x86_64";
and what's in /etc/dhcp/dhclient.d is -rwxr-xr-x. 1 root root 3221 Nov 11 2010 nis.sh -rwxr-xr-x. 1 root root 1944 May 20 2009 ntp.sh
What kind of configuration were you thinking of that could be wrong?
Well, something that requests a "bootfile-name" from the dhcp server. That's not a valid option for ISC dhcpd, but could be for other implementations.
-
Johnny Hughes -
Lars Hecking -
m.roth＠5-cent.us