Skip to content

Conversation

@mselim00
Copy link
Contributor

Issue #, if available:

Description of changes:
The EFA installer comes with a utility to validate the installation by ping ponging traffic over EFA on localhost. This adds a step to run that in the unit test.

Sample test output:

=== RUN   TestPingPong
=== RUN   TestPingPong/pingpong
    pingpong_test.go:155: need at least 2 nodes with EFA capacity, got 1
--- FAIL: TestPingPong (0.07s)
    --- FAIL: TestPingPong/pingpong (0.07s)
=== RUN   TestUnit
=== RUN   TestUnit/unit
=== RUN   TestUnit/unit/Unit_test_succeeds
=== NAME  TestUnit/unit
    unit_test.go:132: Logs for pod "efa-unit-0"
        Running test on a g5.24xlarge
        provider: efa
            fabric: efa-direct
            domain: rdmap0s26-rdm
            version: 201.0
            type: FI_EP_RDM
            protocol: FI_PROTO_EFA
        provider: efa
            fabric: efa
            domain: rdmap0s26-rdm
            version: 201.0
            type: FI_EP_RDM
            protocol: FI_PROTO_EFA
        provider: efa
            fabric: efa
            domain: rdmap0s26-dgrm
            version: 201.0
            type: FI_EP_DGRAM
            protocol: FI_PROTO_EFA
        Verified at least 1 DGRAM endpoint(s) are available (found 1)
        Verified at least 1 RDM endpoint(s) are available (found 2)
        Running single-node efa test
        Localhost fi_pingpong test: Attempt 1 (max 3)...
        Starting server on port 49220...
        Starting client on port 64270...
        Server Log:
        bytes   #sent   #ack     total       time     MB/sec    usec/xfer   Mxfers/sec
        64      10      =10      1.2k        0.00s      0.87      73.20       0.01
        256     10      =10      5k          0.00s     15.61      16.40       0.06
        1k      10      =10      20k         0.00s     61.50      16.65       0.06
        4k      10      =10      80k         0.00s    227.56      18.00       0.06
        Client Log:
        bytes   #sent   #ack     total       time     MB/sec    usec/xfer   Mxfers/sec
        64      10      =10      1.2k        0.00s      0.88      72.65       0.01
        256     10      =10      5k          0.00s     16.15      15.85       0.06
        1k      10      =10      20k         0.00s     63.80      16.05       0.06
        4k      10      =10      80k         0.00s    235.40      17.40       0.06
        fi_pingpong: SUCCESS!
        Success!
--- PASS: TestUnit (10.15s)
    --- PASS: TestUnit/unit (10.15s)
        --- PASS: TestUnit/unit/Unit_test_succeeds (10.01s)
FAIL
FAIL    github.com/aws/aws-k8s-tester/test/cases/efa    31.243s
FAIL

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

@mselim00 mselim00 requested a review from ndbaker1 July 25, 2025 21:43
@mselim00 mselim00 merged commit 41f708c into aws:main Jul 26, 2025
10 checks passed
@mselim00 mselim00 deleted the efa-single branch July 26, 2025 00:12
annapoornanarayan pushed a commit to annapoornanarayan/aws-k8s-tester that referenced this pull request Aug 11, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants