test_cert and test_SubCAkeyReplication are randomly failing.
The suspect for test_SubCAkeyReplication is an nss bug:
https://bugzilla.redhat.com/show_bug.cgi?id=1985061
The reason for test_cert failures was not identified, the only
relevant line in the log contains:
2021-07-22T17:37:21.0873339Z tests: cert, result: 1, time: 30:08.98
2021-07-22T17:37:21.0874172Z Command exited with non-zero status 1
Disable these tests in gating until the NSS bug is fixed and
the related build is available in Fedora.
Related: https://pagure.io/freeipa/issue/8864
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
Reviewed-By: Stanislav Levin <slev@altlinux.org>
A kinit immediately following a password change can fail.
Setting KRB5_TRACE and retrieving kdcinfo will help to understand
the cause of failure.
Fixes: https://pagure.io/freeipa/issue/8929
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
If dynamic DNS updates are selected, sssd will use GSS-TSIG
by default for nsupdate.
When ipa-client-install notices that plain nsupdate is required,
switch sssd to use no authentication for dynamic updates too.
Fixes: https://pagure.io/freeipa/issue/8402
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
ipa-client-install invokes nsupdate with GSS-TSIG at client
enrollment time. If that fails, no retry is done.
Change that behavior to try again without GSS-TSIG.
Fixes: https://pagure.io/freeipa/issue/8402
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
If nsds5replicalastupdateend is not yet present,
ipa-replica-manage will backtrace as it tries to retrieve that
attribute unconditionally.
Gracefully handle that situation.
Fixes: https://pagure.io/freeipa/issue/8605
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
Handle package removal without taking dependencies into account.
E.g. add frontends for rpm -e --nodeps.
Related: ipatests/pytest_ipa/integration/tasks.py
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Armando Neto <abiagion@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
On 10.10+ releases of Dogtag, the PKI installer will not depend
on sudo anymore. This opens the possibility of creating IPA servers
without a properly configured sudo.
In fact, even IPA clients should have sudo and libsss_sudo installed
in most cases, so add a weak dependency on both of them to the client
subpackage.
Also make sure libsss_sudo is installed if sudo is present.
Fixes: https://pagure.io/freeipa/issue/8530
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Armando Neto <abiagion@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
df uses 1024 bytes as its default display value, but this can be
tweaked by environment variables or a CLI knob.
Force the output unit to 1024 bytes using the CLI and parse it
accordingly.
Fixes: https://pagure.io/freeipa/issue/8674
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
Previously, dns_lookup_kdc was only set to True if DNS
discovery worked or if the KDC was not specified on the
command-line.
Setting dns_lookup_kdc to False would result in a hardcoded
configuration which is less reliable in the long run.
For instance, adding a trust to an Active Directory forest
after clients are enrolled would result in clients not being
able to authenticate AD users. Recycling FreeIPA servers
could prove problematic if the original hostnames are not
reused too.
Change summary:
Always set dns_lookup_kdc to True on client enrollment.
With this change, DNS SRV search will always be performed
before looking into /etc/krb5.conf realm entries.
Fixes: https://pagure.io/freeipa/issue/6523
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Previously, dns_lookup_kdc was only set to True if DNS
discovery worked or if the KDC was not specified on the
command-line.
Make sure dns_lookup_kdc is always set to true.
Fixes: https://pagure.io/freeipa/issue/6523
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Use get_kdcinfo before and after kinit if krb5_trace in kinit_as_user.
This will help determine how SSSD was selecting which KRB5KDC to use.
Fixes: https://pagure.io/freeipa/issue/8510
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
Use get_kdcinfo before and after kinit if krb5_trace in create_active_user.
This will help determine how SSSD was selecting which KRB5KDC to use.
Fixes: https://pagure.io/freeipa/issue/8510
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
get_kdcinfo(host) retrieves /var/lib/sss/pubconf/kdcinfo.$REALM on host.
It also logs whether SSSD considers the IPA domain as Online or not before
and after retrieving the file.
Related: https://pagure.io/freeipa/issue/8510
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
Split wait_for_sssd_domain_status_online so that we can easily check
that SSSD considers the IPA domain online with check_if_sssd_is_online.
Related: https://pagure.io/freeipa/issue/8510
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
Enhance test_EPN_nbdays so that it checks:
* that no emails get sent when using --dry-run
* that --from-nbdays implies --dry-run
* that --to-nbdays requires --from-nbdays
* illegal inputs for nbdays:
** from-nbdays > to-nbdays
** non-numerical input
** decimal input
Fixes: https://pagure.io/freeipa/issue/8449
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
Paramiko is not compatible with FIPS.
Migrate test_2fa_enable_single_prompt to the OpenSSH CLI SSH(1).
Rationale: this is exactly what clients use.
Also add a warning when test_2fa_disable_single_prompt is executed in FIPS mode.
Fixes: https://pagure.io/freeipa/issue/8129
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Mohammad Rizwan <myusuf@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
Paramiko is not compatible with FIPS.
Migrate run_cmd_on_ui_host to the OpenSSH CLI SSH(1) using
tasks.py's run_ssh_cmd.
Rationale: this is exactly what clients use.
Fixes: https://pagure.io/freeipa/issue/8129
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Mohammad Rizwan <myusuf@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
Sometimes test_login_wrong_password fails because the log window the
string message is searched in is too narrow.
Broaden the window by looking at the past 10 seconds.
Fixes: https://pagure.io/freeipa/issue/8432
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
Due to a change in authselect, rolling back the installation
does not produce the same nsswitch.conf as on a clean install.
Mark the test xfail until ipa-client-install is enhanced to
use authselect profile backup/restore.
Related: https://pagure.io/freeipa/issue/8189
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
The test test_adtrust_install.py::TestIpaAdTrustInstall::test_add_agent_not_allowed
sometimes fails at kinit in create_active_user:
```
kinit: Password has expired while getting initial credentials
```
krb5_strace shows that this happens when kinit changes servers
between password change and TGT requests.
Display SSSD's kdcinfo to see if kinit should be pinned to one
server.
Related-to: https://pagure.io/freeipa/issue/8353
Related-to: https://pagure.io/freeipa/issue/8271
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
Reviewed-By: Sumit Bose <sbose@redhat.com>
Reviewed-By: Robbie Harwood <rharwood@redhat.com>
EPN stands for Expiring Password Notification. It is a standalone
tool designed to build a list of users whose password would expire
in the near future, and either display the list in a machine-readable
format, or send email notifications to these users.
EPN provides command-line options to display the list of affected users.
This provides data introspection and helps understand how many emails
would be sent for a given day, or a given date range.
The command-line options can also be used by a monitoring system to alert
whenever a number of emails over the SMTP quota would be sent.
EPN is meant to be launched once a day from an IPA client (preferred)
or replica from a systemd timer.
EPN does not keep state. The list of affected users is built at runtime
but never kept.
TLS/STARTTLS SMTP code is untested and unlikely to work as-is.
Parts of code contributed by Rob Crittenden.
Ideas and feedback contributed by Christian Heimes and Michal Polovka.
Fixes: https://pagure.io/freeipa/issue/3687
Signed-off-by: François Cami <fcami@redhat.com>
Signed-off-by: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
The design draft lists the user stories, implementation choices,
implementation details, limitations, and changes for the new
Expired Password Notifications (EPN) feature.
Fixes: https://pagure.io/freeipa/issue/3687
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Use -W with sphinx-build by default to turn warnings into errors.
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Add doclint to fastcheck so that documentation syntax issues
are caught sooner (before they hit CI).
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
* tasks had an ipa_backup() method that was not used anywhere.
* test_backup_and_restore had a backup() method that used to return
both the path to the backup and the whole result from run_command ;
The path to the backup can be determined from the result.
Clean up:
* move test_backup_and_restore.backup to tasks.ipa_backup, replacing
the unused method.
* add tasks.get_backup_dir(host) which runs ipa-backup on host and
returns the path to the backup directory.
* adjust test_backup_and_restore and test_replica_promotion.
Related: https://pagure.io/freeipa/issue/8217
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
test_fips takes between 45 and ~80 mins to run.
The templates' timeout was 3600s which is too short for
successful execution. 7200s should do.
Fixes: https://pagure.io/freeipa/issue/8247
Reviewed-By: Sergey Orlov <sorlov@redhat.com>
ipa-backup does not check whether the IPA master it is running on has
all used roles installed. This can lead into situations where backups
are done on a CAless or KRAless host while these roles are used in the
IPA cluster. These backups cannot be used to restore a complete cluster.
With this change, ipa-backup refuses to execute if the roles installed
on the current host do not match the list of roles used in the cluster.
A --disable-role-check knob is provided to restore the previous behavior.
Fixes: https://pagure.io/freeipa/issue/8217
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Mohammad Rizwan Yusuf <myusuf@redhat.com>
ipa-backup should refuse to execute if the local IPA server does not
have all the roles used in the cluster.
A --disable-role-check knob should also be provided to bypass the
check.
Add an integration test for the new behavior and the knob.
Related: https://pagure.io/freeipa/issue/8217
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Mohammad Rizwan Yusuf <myusuf@redhat.com>
The Hidden replica tests did not test what happened when KRA was
installed on a hidden replica and then other KRAs instantiated from
this original one. Add a test scenario that covers this.
Related: https://pagure.io/freeipa/issue/8240
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Michal Polovka <mpolovka@redhat.com>