Clean up the PKI securitydomain when removing a server

[rcritten]

PKI has its own internal knowledge of servers and services
in its securitydomain. This has not been cleaned up in the
past but is becoming more of an issue as PKI now relies on its
securitydomain for more things, and it has a healthcheck that
reports inconsistencies.

Removing entries is straightforward using the PKI REST API.

In order to operate on the API access is needed. There was an
unused Security Domain Administrators group that I've added to
the resourceACLS we created for managing the securitydomain.
The ipara user is added as a member of this group. The REST
API binds to the CA using the IPA RA certificate.

Related commits are b3c2197
and ba4df64.

These resourceACLS were originally created as a backwards
compatibility mechanism for dogtag v9 and later only created when a
replica was installed purportedly to save a restart. I don't see
any reason to not have these defined. They are apparently needed due
to the PKI database upgrade issues.

In any case if the purpose was to suppress these ACLS it failed
because as soon as a replica with a CA was installed they were as
well, and we need this ACL in order to manage the securitydomain.

https://pagure.io/freeipa/issue/8930

Signed-off-by: Rob Crittenden rcritten@redhat.com

[flo-renaud]

When the topology is CA-less, this fails with:

# ipa server-del replica.ipa.test
Removing replica.ipa.test from replication topology, please wait...
ipa: ERROR: an internal error has occurred

The code must be called only when CA is enabled.

[rcritten]

Fixed and added a CAless test to the temp commit. Nice catch, thanks.

[rcritten]

test_integration/test_server_del.py::TestLastServices::test_forced_removal_of_master due to KDC being unavailable. Not sure why this would be affected, going to re-run it.

[flo-renaud]

The real issue is this one:

INFO     ipatests.pytest_ipa.integration.host.Host.replica1.IPAOpenSSHTransport:transport.py:391 RUN ['ipa', 'server-del', 'master.ipa.test', '--ignore-last-of-role']
DEBUG    ipatests.pytest_ipa.integration.host.Host.replica1.cmd139:transport.py:513 RUN ['ipa', 'server-del', 'master.ipa.test', '--ignore-last-of-role']
DEBUG    ipatests.pytest_ipa.integration.host.Host.replica1.cmd139:transport.py:557 ipa: ERROR: cannot connect to 'https://master.ipa.test:443/ca/rest/account/login': [Errno 111] Connection refused
DEBUG    ipatests.pytest_ipa.integration.host.Host.replica1.cmd139:transport.py:557 Removing master.ipa.test from replication topology, please wait...
DEBUG    ipatests.pytest_ipa.integration.host.Host.replica1.cmd139:transport.py:217 Exit code: 1

If the last CA is force-removed, the server-del probably removes the CA role from the master before it tries to call the CA to update the security domain. I suspect the code is failing because it doesn't know which server to contact for the security domain op.

[rcritten]

Ok. I added a broad try/except. I don't think this should block removal but the user should be notified that the server hasn't been removed from the security domain.

[flo-renaud]

Hi @rcritten
the test_server_del error happened because the previous test did not cleanup properly.

    def test_removal_of_server_raises_error_about_last_kra(self):
        """
        test that removal of server fails on the last KRA

        We shut it down to verify that it can be removed if it failed.
        """
        tasks.install_kra(self.master)
        self.master.run_command(['ipactl', 'stop'])
        tasks.assert_error(
            tasks.run_server_del(self.replicas[0], self.master.hostname),
            "Deleting this server is not allowed as it would leave your "
            "installation without a KRA.",
            1
        )
        # Restarting the server we stopped is not necessary as it will
        # be removed in the next test.

After test_removal_of_server_raises_error_about_last_kra, ipa services are stopped on the master and that explains the failure detected by PRCI. Could you fix this test (stop/restart ipa services in a fixture)?
We will need the try/except modification that you added anyway (for instance if calling server-del for a server that crashed, was decommissioned and was the last CA).

[rcritten]

Huh, that test worked fine locally. I wonder if I specifically just ran that one test instead of the whole suite. I'll add a fixture.

[flo-renaud]

Hi @rcritten
thanks for the updated patch, WFM. You can remove the temp commit and I will ack when prci completes.

[rcritten]

vagrant provisioning errors, re-running.

[flo-renaud]

master:

• db69855 Clean up the PKI securitydomain when removing a server
• c0d6c05 ipatests: Verify that securitydomain is updated on server-del