SecureComms: E2e Test SecureComms without KBS #2089
Conversation
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
from
d7b9676 to
6dcef0f
Compare
davidhadas
changed the title
davidhadas
changed the title
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
25 times, most recently
from
f08e2d3 to
f582a4e
Compare
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
from
b7f3d93 to
815170c
Compare
|
So this change is just regression testing to check that the no_trustee version of securecomms doesn't cause any failures. Is there any easy way to validate that this set-up has been used? |
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
3 times, most recently
from
26d3602 to
524869d
Compare
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
3 times, most recently
from
9b5b9b4 to
0c3db17
Compare
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
3 times, most recently
from
a7cf406 to
2331371
Compare
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
4 times, most recently
from
c89bfc7 to
be301d2
Compare
| if !testCase_secureComms_isActive { | ||
| t.Skip("Skip - SecureComms is configured to be inactive - no need to test") | ||
| } |
There was a problem hiding this comment.
We typically do the skip in the provider test case, so libvirt_test in this case, so it might be worth following this pattern.
Also our current pattern for whether to skip tests or not is to use environment variables, so can you use TEST_E2E_SECURE_COMMS rather than creating this extra go variable?
There was a problem hiding this comment.
Does it really make sense to move the skip to libvirt_test, it needs to be skipped no matter who the provider is...
seems like common_suite is the natural place for it...
There was a problem hiding this comment.
I understand your reasoning, but none of the other providers would call this test, unless you are planning to support testing this elsewhere in future?
There was a problem hiding this comment.
I dont see why SecureComms will not be tested with other providers in the future.
There was a problem hiding this comment.
TEST_E2E_SECURE_COMMS is a matrix parameter. It may have multiple values and the mapping is done once in config_libvirt.sh. It results in potentially activating SecureComms. Here we need to identify if secureComms was activated in this matrix option and act accordingly. This is done by checking if props["SECURE_COMMS"] == "true" as shown on e2e/main_test.go.
Therefore we should not be using TEST_E2E_SECURE_COMMS env directly, but have our own env variable indicating is SecureComms is active - I am now using the SECURE_COMMS env variable for that.
| if props["SECURE_COMMS"] == "true" { | ||
| testCase_secureComms_isActive = true | ||
| log.Info("Do setup secureComms is active") | ||
| } |
There was a problem hiding this comment.
This code would not be needed if you followed the env var approach
src/cloud-api-adaptor/test/provisioner/libvirt/provision_common.go
Outdated
Show resolved
Hide resolved
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
7 times, most recently
from
551140b to
6bd5e10
Compare
Add support for e2e testing SecureComms without KBS Signed-off-by: David Hadas <[email protected]>
davidhadas
force-pushed
the
secComms_test_no_kbs
branch
from
6bd5e10 to
61929c7
Compare
Support testing SecureComms without KBS
See
Should be merged after #2065