[ovs-dev] BUG: Weird failure of OVS classifier

Jean Tourrilhes jean.tourrilhes at hpe.com
Mon Mar 30 23:04:20 UTC 2020


On Sun, Mar 29, 2020 at 07:37:43PM -0700, Ben Pfaff wrote:
> 
> Usually this kind of thing means that userspace and the kernel don't
> have the same idea of the definition of a flow.  Tracking it down is
> often tricky.

	This is what I was kind of guessing. Thanks for the explanation.
	The worrying thing is that I'm not doing anything fancy or
weird with the flow table, this is the full state of the userspace
flow table at the time the bug is triggered :
------------------------------------------------
NXST_FLOW reply (xid=0x4):
 cookie=0x0, duration=985.656s, table=0, n_packets=671126, n_bytes=963599304, idle_age=195, priority=32769,in_port=LOCAL actions=output:1
 cookie=0x0, duration=985.652s, table=0, n_packets=23219, n_bytes=2661008, idle_age=0, priority=32769,in_port=1 actions=LOCAL
------------------------------------------------
	Obviously, the kernel flow table is a bit more complicated...

> This kind of miserable problem is one reason why I wish that OVS could
> switch to an entirely packet-based interface.  The user/kernel interface
> is almost unmaintainable.

	I think this is orthogonal to what I'm seeing.
	Thanks a lot !

	Jean


More information about the dev mailing list