I went into the "bad key length" in below datapath flow in a test
environment:
recirc_id(0x109),in_port(5),ct_state(-new+est-rel-inv+trk),eth(src=fa:16:3e:00:49:66,dst=fa:16:3e:ca:4a:20),eth_type(0x0800),ipv4(src=
22.22.22.4/255.255.255.252,dst=22.22.23.4,tos=0/0x3,ttl=64,frag=no),
packets:4, bytes:392, used:0.740s, actions:set(eth(bad key length 24,
expected 12)(fa 16 3e b8 13 c6 fa 16 3e 2c d3 a2 20 00 0b 00 1c 00 07 00 16
16 16 04)),set(ipv4(src=22.22.22.4/255.255.255.252,dst=22.22.23.4,ttl=63)
),ct(commit),set(tunnel(tun_id=0x8,dst=10.65.241.187,ttl=64,geneve({class=0xffff,type=0,len=4,0x30004}),flags(df|key))),2
The user space flows look good when checking with ofproto/trace:
$ sudo ovs-appctl ofproto/trace
"recirc_id(0x109),in_port(5),ct_state(-new+est-rel-inv+trk),eth(src=fa:16:3e:00:49:66,dst=fa:16:3e:ca:4a:20),eth_type(0x0800),ipv4(src=
22.22.22.4/255.255.255.252,dst=22.22.23.4,tos=0/0x3,ttl=64,frag=no)"
...
Final flow:
recirc_id=0x109,ct_state=est|trk,ip,reg5=0x1,reg6=0x4,reg7=0x3,metadata=0x7,in_port=14,vlan_tci=0x0000,dl_src=fa:16:3e:00:49:66,dl_dst=fa:16:3e:ca:4a:20,nw_src=22.22.22.4,nw_dst=22.22.23.4,nw_proto=0,nw_tos=0,nw_ecn=0,nw_ttl=64
Megaflow:
recirc_id=0x109,ct_state=-new+est-rel-inv+trk,ip,tun_id=0,in_port=14,vlan_tci=0x0000/0x1000,dl_src=fa:16:3e:00:49:66,dl_dst=fa:16:3e:ca:4a:20,nw_src=
22.22.22.4/30,nw_dst=22.22.23.4,nw_ecn=0,nw_ttl=64,nw_frag=no
Datapath actions:
set(eth(src=fa:16:3e:b8:13:c6,dst=fa:16:3e:2c:d3:a2)),ct(commit),set(tunnel(tun_id=0x8,dst=10.65.241.187,ttl=64,geneve({class=0xffff,type=0,len=4,0x30004}),flags(df|key))),2
I don't know what the extra 12 bytes are, but the last 4 bytes is the src
IP: 16 16 16 04 == 22.22.22.4.
However this didn't impact the actual behaviour, and I see it in one
environment only. Is this a known problem? What could be the root cause
here?
--
Best regards,
Han
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
