Improved integration test coverage for farm build functionality.
#26973
Conversation
Existing integration test cases for `farm` functionality
is limited to the operations that create/update/delete farm
objects. There is no test coverage of any actual farm builds.
This submission seeks to address this via the addition of
a new integration test module: `farm_build_test.go`
This module:
- prepares a simulated, multi-farm, multi-node test environment;
- performs testing of the podman farm build function across
that environment;
- confirms that the expected number of build are performed; that
they are of the correct architecture; and that the builds occur
on the expected node, given the build parameters that were supplied.
The commit also includes two minor changes:
- A very minor change the farm documention to clarify one of
of the `default` setting.
- The farm system-test script 001-farm.bats now uses an
explicitly set connection name rather than relying on
the default in the users environment being set in a
supportive manner.
Signed-off-by: kevin <[email protected]>
|
/test ? |
|
@kcrocombe: Cannot trigger testing until a trusted user reviews the PR and leaves an In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
/retest |
|
@kcrocombe: Cannot trigger testing until a trusted user reviews the PR and leaves an In response to this:
Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
|
@containers/podman-maintainers: Is it possible to re-run the above failing checks please? Can't draw a line between the test failures and anything in my PR, so want to rule out flakiness. Also is it possible for me to run re-run these myself? I couldn't work out how to. |
|
it is not possible to rerun yourself and it is quite likely a flake. ive rerun the test now and will try to keep an eye on it. thanks for the submission. |
|
Thanks for doing that. Though with the same result unfortunately. The failing test seems to think an additional google nameserver has been introduced to host/resolv.conf that it wasn't expecting to see. Still struggling to see why a changes to a test-case and a bit of documentation would be responsible for that, though. |
|
Spoke too soon...thanks for sorting that out for me. Much appreciated. |
|
The general approach here (reverse-proxy to simulate multiple nodes) is inspired - nice work! |
|
[APPROVALNOTIFIER] This PR is APPROVED This pull-request has been approved by: kcrocombe, mheon The full list of commands accepted by this bot can be found here. The pull request process is described here
Needs approval from an approver in each of these files:
Approvers can indicate their approval by writing |
openshift-ci
bot
added
the
approved
@mheon Yeah, I thought so too; which is why I stole the idea, without shame, from whoever wrote system_connection_test.go (@ben-krieger possibly). I thank you on his behalf. |
|
@baude : This may have skipped ahead of it's self. Though approved, I understand it still needs a second pair of eyes to move forward. Could you have look when you've got a moment?. Ta. |
|
Losing a bit of traction on this, I think, and its now backing in to the couple of PR's I have to follow on. To keep things moving, I suggest I now pull this PR and wrap its contents into its successor? Please advise me otherwise if this isn't the best plan. Ta. |
|
I'm good to merge as-is. |
Existing integration test cases for
farmfunctionality is limited to the operations that create/update/delete farm objects. There is no test coverage of any actual farm builds.This submission seeks to address this via the addition of a new integration test module:
farm_build_test.goThis module:
The commit also includes two minor changes:
A very minor change the farm documention to clarify one of of the
defaultsetting.The farm system-test script 001-farm.bats now uses an explicitly set connection name rather than relying on the default in the users environment being set in a supportive manner.
Does this PR introduce a user-facing change?