[NETVIRT-525] ELAN Datapath failed with 100% packet loss Created: 13/Mar/17  Updated: 18/Apr/17  Resolved: 18/Apr/17

Status: Resolved
Project: netvirt
Component/s: General
Affects Version/s: Boron
Fix Version/s: None

Type: Bug
Reporter: Suvitha Balu Assignee: P Govinda Rajulu
Resolution: Won't Do Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Operating System: All
Platform: All

External issue ID: 7961

 Description   

Netvirt-ELAN service
The ELAN data-path test failed with 100% packet loss in Sandbox.

I have verified the flow table and it has entry to drop the packet.
Log:
https://logs.opendaylight.org/sandbox/jenkins091/netvirt-csit-1node-openstack-mitaka-upstream-transparent-boron/3

In some build, When i do VM delete and recreate then the flow table updated with corresponding port and data-path works fine.
Log:
https://logs.opendaylight.org/sandbox/jenkins091/netvirt-csit-1node-openstack-mitaka-upstream-learn-boron/1

 Comments   
Comment by govind [ 18/Apr/17 ]

This issue is fixed.

code-review contains fix: https://git.opendaylight.org/gerrit/#/c/53588/

Please test with the latest build.

Comment by Suvitha Balu [ 18/Apr/17 ]

I have tested with latest boron and carbon build and seen the same issue.

carbon log:
https://logs.opendaylight.org/sandbox/jenkins091/netvirt-csit-1node-openstack-newton-upstream-learn-carbon/9

boron log:
https://logs.opendaylight.org/sandbox/jenkins091/netvirt-csit-1node-openstack-newton-upstream-transparent-boron/1

Comment by Periyasamy Palanisamy [ 18/Apr/17 ]

Hi,

On the Carbon report:

Here I see two tunnel interfaces are created b/w DPNs (like below) which is wrong.

{ "destination-DPN": 123772809692053, "source-DPN": 49069236146780, "transport-type": "odl-interface:tunnel-type-vxlan", "tunnel-interface-names": [ "tuna8f4b6ca395", "tun231de31af6e" ] }

There should be only one tunnel interface exist b/w DPNs. Also out of which one tunnel interface doesn’t have any interface state created.
Due to this, when ELAN queries tunnel interface name from ITM and it returns interface name which doesn’t have interface state.
This leads to remote DMAC (51) programmed with drop action and Remote BC group is not updated.

Can you make sure single tunnel interface exist between DPNs ?

On the Boron report:

There is ITM tunnel state up event is received b/w dpIds 18042706445918 and 83764721875373 at 2017-04-18 06:00:02,613 and ELAN is able to retrieve tunnel egress action properly till 2017-04-18 06:57:23,087. Suddenly it stopped working till 2017-04-18 07:07:40,801 which table 51 is programmed with drop action.

Is tunnel got deleted during this period ?

Please make sure ITM tunnels are configured correctly before running ELAN tests across DPNs if you are not using auto-tunnel configuration.

Generated at Wed Feb 07 20:21:47 UTC 2024 using Jira 8.20.10#820010-sha1:ace47f9899e9ee25d7157d59aa17ab06aee30d3d.