[ovs-discuss] Drops in OVS

Ben Pfaff blp at nicira.com
Fri Jul 31 01:22:00 UTC 2015


On Fri, Jul 31, 2015 at 06:48:26AM +0530, Abhishek Verma wrote:
> > OVS wouldn't normally drop packets.  It would normally happen before or
> > after OVS.
> >
> > OVS *can* drop packets at the kernel/user interface.  Those would show
> > up in "ovs-dpctl show" as "lost" packets.
> >
> > Are you looking at the OVS vport stats or the stats via e.g. "ifconfig"?
> > Both are relevant.
> >
> 
> I look at
> 
> root#ovs-vsctl get interface port-to-routerA statistics
> {collisions=0, rx_bytes=2845224091, rx_crc_err=0, *rx_dropped=0*,
> rx_errors=0, rx_frame_err=0, rx_over_err=0, rx_packets=2469921,
> tx_bytes=1390987, *tx_dropped=0*, tx_errors=0, tx_packets=11200}
> root#
> 
> and the ifconfig statistics.
> 
> root# ovs-dpctl show
> system at ovs-system:
> lookups: hit:50254324 missed:15221 *lost:4424*
> flows: 0
> port 0: ovs-system (internal)
> port 1: br0 (internal)
> port 2: port-to-routerA (gre: df_default=false, ttl=0)
> root#
> 
> I just see *4424 *packets getting lost. What are these packets? Any idea
> when would we lose them?

"lost" means that a queue from the kernel to userspace overflowed.  It
would ordinarily happen if there was a big burst of packets that didn't
match in the kernel flow table.  Packets can also be lost this way if
OVS is stopped and restarted.

Since this is a GRE tunnel, packets can get lost in the kernel TCP/IP
stack GRE processing on ingress or egress, and that wouldn't show up in
any of these places.



More information about the discuss mailing list