Srikanth,
The reason wget works and doesnt send packets to the VM is because your
classification rule (pasted below) matches on srcPort=2000 and
dstPort=80, but you dont specify the srcPort with wget, so the packets
dont match classification, and subsequently dont get sent to SFC (hence
they dont go the VNF VM). The packets go directly from the client to the
server.
This is your classification flow:
cookie=0x1110010000510255, duration=1268.611s, table=11,
n_packets=896, n_bytes=66304, tcp,reg0=0x1,tp_src=2000,tp_dst=80
actions=move:NXM_NX_TUN_ID[0..31]->NXM_NX_NSH_C2[],push_nsh,load:0x1->NXM_NX_NSH_MDTYPE[],load:0x3->NXM_NX_NSH_NP[],load:0xc0a8001a->NXM_NX_NSH_C1[],load:0x33->NXM_NX_NSP[0..23],
load:0xff->NXM_NX_NSI[],load:0x7b7b7b03->NXM_NX_TUN_IPV4_DST[],load:0x33->NXM_NX_TUN_ID[0..31],resubmit(,0)
Looking at the flows in more detail, I dont see the SFC flows, which is
why the packets dont go to the VNF VM when you use curl. When SFC is
used with Netvirt for OPNFV (which is what you're doing here) the SFC
flows should be in tables 0 and tables 152-158. The largest table I see
here is 111. Netvirt is working correctly, and sending the packets to
SFC, but the SFC flows dont exist.
Can you check the ODL logs to see if something failed in SFC. The ODL
logs are usually in /opt/opendaylight/data/logs/karaf.log* Just do this
grep to see what happened:
grep -i sfc /opt/opendaylight/data/logs/karaf.log*
Regards,
Brady
On 07/02/17 09:24, Srikanth Lingala wrote:
Hi,
My setup includes:
·One Openstack Controller with ODL (of course with SFC) which is
deployed through OPNFV Colorado 3.0
·One aarch64 Compute Node (OpenStack Mitaka), which is attached to the
above OS Controller
I downloaded OVS 2.6.1 from OVS git hub and applied NSH patches from
below URL:
https://github.com/yyang13/ovs_nsh_patches/tree/master/v2.6.1
I am able to create the SFC attributes like VNFD, VNF, Chain, and
Classifier through Tim Rozet SFC walkthrough
(https://github.com/trozet/sfc-random/blob/master/tacker_sfc_apex_walkthrough.txt).
And also, I launched VNF, http_server and http_client VM’s on the same
compute node.
All the related NSH flows are added to the bridge ‘br-int’ in the
compute node without fail.
But, when I execute the command ‘*curl --local-port 2000
123.123.123.4*’ from the http_client VM, I am getting the below error
message:
*curl: (7) Failed to connect to 123.123.123.4 port 80: Connection
timed out*
When I execute the command ‘*wget 123.123.123.4*’ from the http_client
VM, I am getting the 200 OK response.
But, in both the cases, no VxLAN packets are coming at the tacker VNF
VM (using vxlan_tool.py).
OVS bridges and flows details of Compute Node are given at the below link:
http://pastebin.com/MEN7dk8n
Can anyone give me some clue, to debug the issue.
I also want to know, whether anyone one are succeeded while executing
SFC across compute node through ODL and OVS 2.6.1 with NSH patches
from Yang.
Thanks for the help.
Regards,
Srikanth.
_______________________________________________
sfc-dev mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/sfc-dev
