[PATCH 0/2] Send agent options on failover bind update

[David W. Hankins]

note; you want dhcp-workers for ISC DHCP related development
discussion, not dhcp-hackers (generic DHCP protocol development
related discussion).

could you forward both patches to dhcp-suggest at isc.org to open a
ticket?  i was going to try and sneak this into 4.2 myself, but now it
looks like i'll only have to do review. :)

we're also looking at moving 'binding scopes' over the failover
channel if that's also of interest to you.

i don't think we've ever run dhcpd through valgrind.  i'd do 4.1.1b1
or wait for 4.1.1b2 just to make it easier and send any results to
dhcp-bugs at isc.org to open a ticket.  those are most up to date with
maintenance atm.

On Fri, Jul 10, 2009 at 11:59:02AM +0200, Sytse Wielinga wrote:
> Hi,
> 
> We're running a set of dhcp servers that depend on relay agent options
> on their leases here, and in that configuration the current dhcpd
> fails to match the client to the lease if it is sent to the 'wrong'
> failover host.
> 
> This patch fixes that by sending the agent options over to the
> failover peers when doing a bind update.
> 
> The patch applies to dhcpd 3.1.2, but I think porting to dhcpd 4
> should be trivial.
> 
> On a sidenote, I ran dhcpd 3.1.1 (debian lenny version) in valgrind,
> and I get the following (unrelated) error on the primary failover
> host:
> 
> ==9639== 1 errors in context 1 of 1:
> ==9639== Invalid read of size 8
> ==9639==    at 0x479987: omapi_one_dispatch (dispatch.c:401)
> ==9639==    by 0x443157: dispatch (dispatch.c:94)
> ==9639==    by 0x403DC4: main (dhcpd.c:608)
> ==9639==  Address 0x54066f0 is 32 bytes inside a block of size 80 free'd
> ==9639==    at 0x4C2130F: free (vg_replace_malloc.c:323)
> ==9639==    by 0x473DD4: dfree (alloc.c:202)
> ==9639==    by 0x474310: omapi_object_dereference (alloc.c:702)
> ==9639==    by 0x478863: omapi_io_dereference (dispatch.c:40)
> ==9639==    by 0x478C39: omapi_unregister_io_object (dispatch.c:138)
> ==9639==    by 0x4767A2: omapi_connection_connect_internal (connection.c:682)
> ==9639==    by 0x47648E: omapi_connection_connect (connection.c:583)
> ==9639==    by 0x479969: omapi_one_dispatch (dispatch.c:418)
> ==9639==    by 0x443157: dispatch (dispatch.c:94)
> ==9639==    by 0x403DC4: main (dhcpd.c:608)
> 
> Doesn't seem trivial to fix though. If you wish I can look into it a
> little bit better.
> 
> Thanks,
> Sytse
> 
> _______________________________________________
> dhcp-hackers mailing list
> dhcp-hackers at lists.isc.org
> https://lists.isc.org/mailman/listinfo/dhcp-hackers
> 

-- 
David W. Hankins	"If you don't do it right the first time,
Software Engineer		     you'll just have to do it again."
Internet Systems Consortium, Inc.		-- Jack T. Hankins
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: <https://lists.isc.org/pipermail/dhcp-hackers/attachments/20090710/672758f2/attachment.bin>