Hi Cathy,
Thanks for your reply. I have the setup done without any errors with only one
vm in the chain. I want to move all the icmp traffic from vm1 to vm3 via vm2.
My Flow classifier looks like:
"neutron flow-classifier-create --ethertype IPv4 --source-ip-prefix
10.0.0.18/32 --destination-ip-prefix 10.0.0.6/32 --protocol icmp FC1"
But using tcpdump on vm2 ingress port, I could not see any traffic. Please let
me know how can I debug this and what could be the possible issue.
Best Regards,
Navdeep Uniyal
From: Cathy Zhang [mailto:cathy.h.zh...@huawei.com]
Sent: Dienstag, 11. Oktober 2016 19:50
To: OpenStack Development Mailing List (not for usage questions)
Subject: Re: [openstack-dev] [networking-sfc][devstack][mitaka]
Hi Navdeep,
Please see inline.
Cathy
From: Navdeep Uniyal [mailto:navdeep.uni...@neclab.eu]
Sent: Tuesday, October 11, 2016 5:42 AM
To: openstack-dev@lists.openstack.org<mailto:openstack-dev@lists.openstack.org>
Subject: [openstack-dev] [networking-sfc][devstack][mitaka]
Hi all,
I have been trying out networking-sfc to create service function chain in
Openstack. I could create all the port pairs, port-pair-groups, flow classifier
and the chain but I could not see the packets on the desired hops.
I am trying to create a simple sfc with 3 VMs(vm1 to vm3) in the setup. I just
want to check how it works. In my setup, vm1 is the Traffic generator(iperf)
and vm3 is the traffic receiver(iperf server). Now, the 2 vms (vm2 and 3) are
in the same network with vm1 and I want to move the iperf traffic from
vm1->vm2->vm3. In order to achieve this, I have created 2 port pairs of vm2
and vm3 and both pairs are in separate port pair groups (PG1 and PG2), also
created a Flow classifier FC1 and finally chain with PG1, PG2 and FC1. Now my
question is, is my setup correct in order to achieve the sfc result as I stated
above? Do I need to include the vm1 in the port pair group?
Cathy> You only need to include VM2 in a port pair group. Traffic source and
traffic destination do not need to be included in the chain's port pair group,
instead their IP addresses should be included in the flow classifier so that
the system knows which flow needs to go through the chain. Here is a link to
thw wiki.
https://wiki.openstack.org/wiki/Neutron/ServiceInsertionAndChaining
Cathy
Below is the flow classifier:
++--+
| Field | Value
|
++--+
| description |
|
| destination_ip_prefix | |
| destination_port_range_max | |
| destination_port_range_min | |
| ethertype| IPv4
|
| id |
e5000ade-50ad-41ed-a159-b89c4blp97ec |
| l7_parameters | {}
|
| logical_destination_port | |
| logical_source_port | 63cdf664-dd67-455c-8345-f01ef58c23e5 |
| name| FC1
|
| project_id |
6b90cd3356144681b44274d4881c5fc7 |
| protocol | tcp
|
| source_ip_prefix | 10.0.0.18/32
|
| source_port_range_max | |
| source_port_range_min | |
| tenant_id |
6b90cd3310104681b44274d4881c5fc7 |
++--+
Is there any wiki with some example case explained with testing scenario?
Best Regards,
Navdeep Uniyal
Email: navdeep.uni...@neclab.eu<mailto:navdeep.uni...@neclab.eu>
-
Software Engineer
NEC Europe Ltd.
NEC Laboratories Europe
Kurfürstenanlage 36, D-69115 Heidelberg,
NEC Europe Ltd | Registered Office: Athene, Odyssey Business Park, West End
Road, London, HA4 6QE, GB | Registered in England 2832014
__
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
