On 10/19/07, Sebastian Reitenbach <[EMAIL PROTECTED]> wrote:
> Hi,
>
> after getting the port ready to compile and install smoothly on OpenBSD,
> meanwhile also tested on sparc ;)
> I recognized that I have problems to build a cluster with more than two
> nodes. First I decided to use broadcast, with a ha.cf file like this:
[skip]
> Oct 19 17:27:01 defiant heartbeat: [10097]: ERROR: write failure on bcast
> fxp0.: Message too long
> Oct 19 17:27:01 defiant heartbeat: [10097]: ERROR: glib: Unable to send
> bcast [-1] packet(len=1543): Message too long
'bcast' is the source of your problem. Accordingly to the
http://www.linux-ha.org/ha.cf#bcast
<cite>
On CRM-enabled clusters, the bcast directive does not work on FreeBSD
because of the fragmentation issue described in
W. Richard Stevens - Unix Network Programming - Vol 1 - 3rd Edition:
The Sockets Networking API
20.4 dg_cli Function Using Broadcasting
..
IP Fragmentation and Broadcasts
Berkeley-derived kernels do not allow a broadcast datagram to be fragmented.
If the size of an IP datagram that is being sent to a broadcast address exceeds
the outgoing interface MTU, EMSGSIZE is returned (pp. 233-234 of TCPv2).
This is a policy decision that has existed since 4.2BSD. There is nothing that
prevents a kernel from fragmenting a broadcast datagram, but the feeling is
that broadcasting puts enough load on the network as it is, so there is no need
to multiply this load by the number of fragments.
....
AIX, FreeBSD, and MacOS implement this limitation. Linux, Solaris, and HP-UX
fragment datagrams sent to a broadcast address.
This results because CRM clusters try and send large (>MTU size)
packets over the cluster communication media.
</cite>
--
Andrew W. Nosenko <[EMAIL PROTECTED]>
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
