Skip any retries and timeout after 1s for model dumps

this is to avoid the 20s timeout we hit in a failed cluster sometimes.

Change-Id: I4e3850a460e107f2e76365fe655dacfae1f5da22
Signed-off-by: Jamo Luhrsen <jluhrsen@redhat.com>

diff --git a/csit/libraries/Netvirt.robot b/csit/libraries/Netvirt.robot
index 90b935af5b346f1e225d816565c065d517545d3a..73ec0bee9a77681b3869d82eff8ab2af7cc7bc0d 100644 (file)
--- a/csit/libraries/Netvirt.robot
+++ b/csit/libraries/Netvirt.robot
@@ -6,7 +6,12 @@ Variables         ../variables/netvirt/Modules.py
 Get Model Dump
     [Arguments]    ${controller_ip}
     [Documentation]    Will output a list of mdsal models using ${data_models} list
-    Create Session    model_dump_session    http://${controller_ip}:${RESTCONFPORT}    auth=${AUTH}    headers=${HEADERS}
+    # while feature request in bug 7892 is not done, we will quickly timeout and not retry the model dump get
+    # request. This is because when it's done in a failed cluster state, it could take 20s for the reesponse to
+    # to come back as the internal clustering times out waiting for a leader which will not come. When bug 7892
+    # is resolved, we can remove the timeout=1 and max_retries=0, but likely have to modify the request itself to
+    # pass a timeout to restconf
+    Create Session    model_dump_session    http://${controller_ip}:${RESTCONFPORT}    auth=${AUTH}    headers=${HEADERS}    timeout=1    max_retries=0
     : FOR    ${model}    IN    @{data_models}
     \    ${resp}=    RequestsLibrary.Get Request    model_dump_session    restconf/${model}
     \    Log    ${resp.status_code}

diff --git a/csit/libraries/OpenStackOperations.robot b/csit/libraries/OpenStackOperations.robot
index 2ab7d7db4185e68c963a1a3e02af82c43a77b334..22f30dff013a1612a7e4d2c9660c484dfb675bbd 100644 (file)
--- a/csit/libraries/OpenStackOperations.robot
+++ b/csit/libraries/OpenStackOperations.robot
@@ -578,12 +578,6 @@ Get OvsDebugInfo
     Run Keyword If    2 < ${NUM_OS_SYSTEM}    Get DumpFlows And Ovsconfig    ${OS_COMPUTE_2_IP}
 
 Get Test Teardown Debugs
-    [Arguments]    ${test_name}=${TEST_NAME}
-    Get OvsDebugInfo
-    Get Model Dump    ${HA_PROXY_IP}
-    Get Karaf Log Events From Test Start    ${test_name}
-
-Get Test Teardown Debugs Allow Model Dump Failures
     [Arguments]    ${test_name}=${TEST_NAME}
     Get OvsDebugInfo
     Run Keyword And Ignore Error    Get Model Dump    ${HA_PROXY_IP}

diff --git a/csit/suites/openstack/clustering/01_l2_tests.robot b/csit/suites/openstack/clustering/01_l2_tests.robot
index ac2f3c62a658e5a5ddaef931a91b27af4071c54d..cdb9720ebf1e06e5ccaea68a002f0cbdf43c65d9 100644 (file)
--- a/csit/suites/openstack/clustering/01_l2_tests.robot
+++ b/csit/suites/openstack/clustering/01_l2_tests.robot
@@ -76,7 +76,6 @@ Delete the Bridge Manually and Verify After Fail
 Bring Up ODL1
     [Documentation]    Bring up ODL1 again
     ClusterManagement.Start Single Member    1
-    [Teardown]    Get Test Teardown Debugs Allow Model Dump Failures
 
 Create Bridge Manually and Verify After Recover
     [Documentation]    Create bridge with OVS command and verify it gets applied from all instances.
@@ -132,7 +131,6 @@ Check Vm Instances Have Ip Address
 Bring Up ODL2
     [Documentation]    Bring up ODL2 again
     ClusterManagement.Start Single Member    2
-    [Teardown]    Get Test Teardown Debugs Allow Model Dump Failures
 
 Ping Vm Instance1 In l2_net_1
     [Documentation]    Check reachability of vm instances by pinging to them.
@@ -183,7 +181,6 @@ Connectivity Tests From Vm Instance3 In l2_net_1
 Bring Up ODL3
     [Documentation]    Bring up ODL3 again
     ClusterManagement.Start Single Member    3
-    [Teardown]    Get Test Teardown Debugs Allow Model Dump Failures
 
 Take Down ODL1 and ODL2
     [Documentation]    Kill the karaf in First and Second Controller
@@ -204,7 +201,6 @@ Connectivity Tests From Vm Instance3 In l2_net_2
 Bring Up ODL1 and ODL2
     [Documentation]    Bring up ODL1 and ODL2 again.
     ClusterManagement.Start Members From List Or All    ${cluster_down_list}
-    [Teardown]    Get Test Teardown Debugs Allow Model Dump Failures
 
 Delete Vm Instance
     [Documentation]    Delete Vm instances using instance names.