Tests: support subid ranges managed by FreeIPA

[aborah-sudo]

issue: #5197

bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1803943

[alexey-tikhonov]

This awaits finalization of RedHatQE/shadow-utils#1 first.

[alexey-tikhonov]

Hi @aborah-sudo, is this ready for review?

[aborah-sudo]

Hi @aborah-sudo, is thi ready for review?

yes

[ikerexxe]

There are some changes to be done but generally it seems good.

By the way, have you ported all the possible improvements that were done in the shadow-utils PR review?

[ikerexxe]

LGTM. Thanks for your work.

[alexey-tikhonov]

@aborah-sudo, is this test executed in PR CI?
At the moment I don't understand how both methods can succeed, because it seems depending on execution order either 'nsswitch.conf::subid' isn't set or IPA subid range isn't generated.

[aborah-sudo]

@aborah-sudo, is this test executed in PR CI?
At the moment I don't understand how both methods can succeed, because it seems depending on execution order either 'nsswitch.conf::subid' isn't set or IPA subid range isn't generated.

the fixture "create_bkp_for_subid_files" will modify the /etc/nsswitch.conf file before every test function

[alexey-tikhonov]

Ok, thanks. Looks much better. ACK.

Please run latest version in idm-ci.

I will leave this to @sgoveas for a final ACK.

[aborah-sudo]

This PR is pass again in IDM-CI :

https://ci-jenkins-csb-idmops.apps.ocp-c1.prod.psi.redhat.com/job/trigger-test-suite-tool/558/console

[alexey-tikhonov]

./src/tests/multihost/ipa/test_subid_ranges.py:126:1: W391 blank line at end of file
./src/tests/multihost/ipa/conftest.py:180:80: E501 line too long (82 > 79 characters)

2. Why tier1 was changed to tier2?

3. why exit_status check was changed to:

        for i in exit_status.readlines():
            assert "write to uid_map failed" not in i

?

[aborah-sudo]

./src/tests/multihost/ipa/test_subid_ranges.py:126:1: W391 blank line at end of file      ---- This is fixed now 
./src/tests/multihost/ipa/conftest.py:180:80: E501 line too long (82 > 79 characters)

1. Why `tier1` was changed to `tier2`?   ----   Its steeves idea to keep IPA tests in tire2 

2. why `exit_status` check was changed to:    ----  IDM ci does not work as we expect .  

exit_status was catching some weird errors which is not related to this test at all:

Check test run with empty list: https://ci-jenkins-csb-idmops.apps.ocp-c1.prod.psi.redhat.com/job/trigger-test-suite-tool/556/console ---- ['Could not chdir to home directory /home/admin: No such file or directory\n'] which not relevant and we need to check newuidmap relate errors which is not

The error we are looking for is : "write to uid_map failed" that is the exact error now configured with the test script . And after puting the exact error which is not occurring at the run time . The test run in IDM CI passed

With exact error code: https://ci-jenkins-csb-idmops.apps.ocp-c1.prod.psi.redhat.com/job/trigger-test-suite-tool/558/console

        for i in exit_status.readlines():
            assert "write to uid_map failed" not in i

?

[alexey-tikhonov]

Frankly, I don't understand what exit_status captures, but (besides one mistype - see recent comment) in general test looks good to me.
@sgoveas , please take care of this further (final review).

[aborah-sudo]

@sgoveas IDMCI passed with this PR without external repo . Every thing is working as expected :

https://ci-jenkins-csb-idmops.apps.ocp-c1.prod.psi.redhat.com/job/trigger-test-suite-tool/724/console

[sgoveas]

Looks good to me

[pbrezina]

Pushed PR: #5755

• master
  • 8e22258 - Tests: support subid ranges managed by FreeIPA