dhcpd -6: buffer overflow detected

perl-list perl-list at network1.net
Fri Mar 30 12:51:26 UTC 2012

We use 4.2.3-p1 built from source on generic (bootstrapped) linux. We notice that it crashes more often than not when working with a renewal packet (I think only renewals) in DHCPv6. I believe that I saw somewhere that ISC is aware of the problem and is working on a solution. To be clear, we do not get any buffer overflow messages in the syslog. 

----- Original Message -----

> From: "Fred Zwarts (KVI)" <F.Zwarts at KVI.nl>
> To: dhcp-users at lists.isc.org
> Sent: Friday, March 30, 2012 6:21:06 AM
> Subject: dhcpd -6: buffer overflow detected

> We are using the ISC dhcpd server for both IPv4 and IPv6 for more
> than a
> year now. Up to recently, the Suse distribution included only version
> 3 of
> isc dhcp, so, in order to run in IPv6 mode, we downloaded the isc
> dhcp
> source code and built the server ourselves.
> We worked around a few problems for IPv6 and finally the
> dhcpd-4.2.1-P1
> version was running in a stable mode for many months now on a Suse
> Linux
> Enterprise System 11 SP1 x86_64 system, both for IPv4 and IPv6.
> Yesterday, we installed Service Pack 2 for SLES11, which now contains
> isc
> dhcpd 4.2.3-P2.
> In IPv4 mode, it was no problem to run the new dhcpd server with
> exactly the
> same configuration as before.
> However, in IPv6 mode the dhcpd server runs for something like one
> minute
> and then stops. In the syslog we find the message:

> Mar 29 13:30:10 SRV014 dhcpd: *** buffer overflow detected ***:
> /usr/sbin/dhcpd6 terminated

> We have tried several time to run the dhcpd server, but each time it
> crashes
> in the same way after about one minute. This happens also when the
> server
> runs in foreground mode.

> Fortunately, we still have the executable of the 4.2.1-P1, so we
> changed
> the soft link /usr/sbin/dhcpd6 to point to this old image. Now the
> server
> does not crash anymore.

> We don't know how exactly the 4.2.3-P2 version was built by Suse. So
> we are
> not sure whether this problem is caused by the way the source code
> was
> compiled, or by a change introduce between the releases of these two
> versions.
> Also we don't know whether there is a specific part in our
> configuration, or
> a specific packet sent by a client that triggers the crash.
> So, my question is whether someone else uses 4.2.3-P2 in IPv6 mode
> without
> problems.
> Also suggestions for diagnosing the problem further are welcome.

> Fred.Zwarts.

> _______________________________________________
> dhcp-users mailing list
> dhcp-users at lists.isc.org
> https://lists.isc.org/mailman/listinfo/dhcp-users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20120330/22a361bb/attachment.html>

More information about the dhcp-users mailing list