[dualtor][orchagent] Add mac move testcases#3395
Conversation
theasianpianist
left a comment
theasianpianist
left a comment
There was a problem hiding this comment.
As per the test plan, I believe these two test cases should be combined into one case - we also want to test the transition between standby and active, and back to standby to make sure the traffic goes to the tunnel/neighbor/tunnel accordingly.
@prsunny, could you please provide some of your thoughts here about how to organize those mac move testcases? |
Thats right. The test is expected as a sequence of operation to perform mac-move and verify. It is better to have it combined in one case. If we have it separated, then maintaining mac-move sequence would be harder. I believe, it will make the test simple as well. |
lolyu
force-pushed
the
add_mac_move
branch
2 times, most recently
from
739a5ac to
070fa6c
Compare
|
@theasianpianist, updated, could you pls take another review? |
Add `test_mac_move` to cover the following scenarios: 1. announce a new neighbor to an active port, then test the traffic. 2. mac move to a standby port, then test the traffic. 3. mac move to another active port, then test the traffic. Signed-off-by: Longxiang Lyu <[email protected]>
|
I don't see the 'Simulate FDB ageout/flush' steps from the test plan, am I missing something? |
FDB will learn from heartbeats immediately after FDB flush/ageout, and stopping the |
Agree with @lolyu that we cannot stop icmp responder. But discussed another method to use a different mac than the interface mac so that it won't get relearnt after fdbflush. @lolyu is trying out the change. |
Signed-off-by: Longxiang Lyu <[email protected]>
|
|
Thank you! |
What is the motivation for this PR? Verify those mac move scenarios: if a new neighbor(in the VLAN subnet) is learned on an active port, the traffic to the new neighbor will be directly forwarded to the neighbor. if the mac moves from an active port to another standby port, verify the traffic is directed back to T1s to the active ToR. if the mac moves back to another active port, verify the traffic is directly forwarded. Signed-off-by: Longxiang Lyu [email protected] How did you do it? Add: test_orchagent_mac_move.py::test_mac_move
We are migrating from Jenkins to azure pipeline. This PR added azure pipeline yaml files and dependent template files for nightly tests. Pipeline yaml files only added for 3 testbeds yet. The pipelines were originally added to branch azp-test of repo https://dev.azure.com/mssonic/internal/_git/sonic-mgmt-int. They have been tested on Azure DevOps. Now we formally add these pipelines to the Networking-acs-sonic-mgmt repo. Currently the internal branch of Networking-acs-sonic-mgmt is synched to same internal branch of the sonic-mgmt-int repo. After this PR is merged, we need to update configuration of the created pipelines to formally use these files from the internal branch. If the pipelines work fine with yaml files from internal branch, we will create more pipeline files based on current templates for rest of the nightly testbeds. Related work items: sonic-net#3021, sonic-net#3073, sonic-net#3135, sonic-net#3153, sonic-net#3162, sonic-net#3176, sonic-net#3238, sonic-net#3241, sonic-net#3346, sonic-net#3352, sonic-net#3378, sonic-net#3389, sonic-net#3395, sonic-net#3397, sonic-net#3398, sonic-net#3407, sonic-net#3410, sonic-net#3411, sonic-net#3412, sonic-net#3413, sonic-net#3414, sonic-net#3415, sonic-net#3434, sonic-net#3437, sonic-net#3445, sonic-net#3446, sonic-net#3447, #9740131, #9821349
What is the motivation for this PR? Verify those mac move scenarios: if a new neighbor(in the VLAN subnet) is learned on an active port, the traffic to the new neighbor will be directly forwarded to the neighbor. if the mac moves from an active port to another standby port, verify the traffic is directed back to T1s to the active ToR. if the mac moves back to another active port, verify the traffic is directly forwarded. Signed-off-by: Longxiang Lyu [email protected] How did you do it? Add: test_orchagent_mac_move.py::test_mac_move
3 participants
Description of PR
Summary:
Fixes #3276
Type of change
Approach
What is the motivation for this PR?
Verify those mac move scenarios:
Signed-off-by: Longxiang Lyu [email protected]
How did you do it?
Add:
test_mac_move.py::test_mac_moveHow did you verify/test it?
Any platform specific information?
Supported testbed topology if it's a new test case?
Documentation