[ofa-general] IPoIB arps disappearing

[Eli Cohen]

On Thu, Jul 10, 2008 at 09:49:54AM -0400, Michael Di Domenico wrote:
> 
> > tests using datagram IPoIB (non-connected mode)
> > dmesg from the compute node
> > cfd-cnsl-0001:~ # dmesg
> > ib0: Send unicast ARP to 0518
> > ib0: Send unicast ARP to 0518
> > ib0: Send unicast ARP to 0518
> > ib0: Send unicast ARP to 0518
> > ib0: Start path record lookup for fe80:0000:0000:0000:00e0:8111:0100:007d
> > MTU > 1024
> > ib0: PathRec LID 0x0161 for GID fe80:0000:0000:0000:00e0:8111:0100:007d
> > ib0: Created ah ffff81042063dc80
> > ib0: created address handle ffff8102206144c0 for LID 0x0161, SL 0
> > ib0: Send unicast ARP to 0161
> > ib0: Send unicast ARP to 0161
> > ib0: Send unicast ARP to 0161
> > ib0: Send unicast ARP to 0518
> > ib0: Send unicast ARP to 0518
> > ib0: Send unicast ARP to 0518
> > ib0: Send unicast ARP to 0161
> > ib0: neigh_destructor for bonding device: ib0
> > ib0: neigh_cleanup for 000404 fe80:0000:0000:0000:00e0:8111:0100:0091
Here I don't see any problem.

> >
> >
> > dmesg from the IO node
> >
> > [root at cfd-io-0001 ~]# dmesg
> > ib0: Send unicast ARP to 0384
> > ib0: Send unicast ARP to 0384
> > ib_mthca 0000:07:00.0: too many gathers
> > ib0: post_send failed
> > ib_mthca 0000:07:00.0: opcode invalid
> > ib0: post_send failed
This looks like a problem that needs investigation but this is not
ConnectX. Are you sure you are using only ConnectX?