Displaying "Dynamic Update" and "Bind update policy" by default
when 'ipa dnszone-show/find' are used would make client dns update
failures easier to diagnose, so display them.
Fixes: https://pagure.io/freeipa/issue/7938
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
When 'ipa krbtpolicy-reset' is called, we need to reset all policy
settings, including per-indicator ones. Per-indicator policy uses
subtyped attributes (foo;bar), the current krbtpolicy-reset code does
not deal with those.
Add support for per-indicator policy reset. It is a bit tricky, as we
need to drop the values to defaults but avoid adding non-per-indicator
variants of the same attributes.
Add test to check that policy has been resetted by observing a new
Kerberos TGT for the user after its policy reset.
Fixes: https://pagure.io/freeipa/issue/8153
Signed-off-by: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
The change to allow overlapping zone to be from the master itself has
introduced two issues: The check for the master itself should only executed
if options.force and options.allow_zone_overlap are both false and the
reverse zone check later on was still handling ValueError instead of
dnsutil.DNSZoneAlreadyExists.
Both issues have been fixed and the deployment with existing name servers
is properly working again.
Fixes: https://pagure.io/freeipa/issue/8150
Signed-off-by: Thomas Woerner <twoerner@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
When re-running `ipa-server-install --setup-dns` on already installed
server, we do not get to the check of being already installed because
DNS zone overlap forces us to fail earlier.
Change exception returned for this case from check_zone_overlap() to
return structured information that allows to understand whether we are
finding a conflict with ourselves.
Use the returned information to only fail DNS check at this point if DNS
zone overlap is generated by a different name server than ourselves.
Reviewed-By: Christian Heimes <cheimes@redhat.com>
DomainValidator.get_trusted_domain_object_from_sid() was using
escape_filter_chars() with bytes. The function only works with text.
This caused idview to fail under some circumstances. Reimplement
backslash hex quoting for bytes.
Fixes: https://pagure.io/freeipa/issue/7958
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
After adding a replica to AD trust agent, the warning
message does not mention that restarting sssd is mantatory
for the trust agent to work. Fix the string.
Fixes: https://pagure.io/freeipa/issue/8148
Signed-off-by: François Cami <fcami@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
ipa-server-install fails if idstart is set to 0. There might be
additional issues when idstart overlaps with local users. Ensure that
idstart is larger than UID_MAX or GID_MAX from /etc/login.defs.
Fixes: https://pagure.io/freeipa/issue/8137
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
The method has_upg returns if user private groups are enabled or
disabled. has_upg() is called three times by user-add. The setting is
now cached on the request local variable context to speed up batch
processing of user imports.
context is cleared after every request.
Related: https://pagure.io/freeipa/issue/8134
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
After an upgrade, the group cn=adtrust agents may be missing some members.
Each ad trust controller must appear twice as member:
- krbprincipalname=cifs/hostname@realm,cn=services,cn=accounts,basedn
- fqdn=hostname,cn=computers,cn=accounts,basedn
Add an upgrade plugin that builds a list of hostnames from the cifs
principals and adds if needed fqdn=hostname...
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1778777
Signed-off-by: Florence Blanc-Renaud <flo@redhat.com>
Reviewed-By: Alexander Bokovoy <abbra@users.noreply.github.com>
urllib3 now supports post-handshake authentication with TLS 1.3. Enable
TLS 1.3 support for Apache HTTPd.
The update depends on bug fixes for TLS 1.3 PHA support in urllib3 and
Apache HTTPd. New builds are available in freeipa-master COPR and in
F30/F31.
Overwrite crypto-policy on Fedora only. Fedora 31 and earlier have TLS
1.0 and 1.1 still enabled by default.
Fixes: https://pagure.io/freeipa/issue/8125
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Thomas Woerner <twoerner@redhat.com>
A raw batch request was fully logged which could expose parameters
we don't want logged, like passwords.
Override _repr_iter to use the individual commands to log the
values so that values are properly obscured.
In case of errors log the full value on when the server is in
debug mode.
Reported by Jamison Bennett from Cloudera
Signed-off-by: Rob Crittenden <rcritten@redhat.com>
Reviewed-by: Florence Blanc-Renaud <frenaud@redhat.com>
For the authentication indicators 'otp', 'radius', 'pkinit', and
'hardened', allow specifying maximum ticket life and maximum renewable
age in Kerberos ticket policy.
The policy extensions are now loaded when a Kerberos principal data is
requested by the KDC and evaluated in AS_REQ KDC policy check. If one of
the authentication indicators mentioned above is present in the AS_REQ,
corresponding policy is applied to the ticket.
Related: https://pagure.io/freeipa/issue/8001
Signed-off-by: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
LDAP attribute options aren't enforced in the LDAP schema. They
represent server- and client-side treatment of attribute values but the
schema definition knows nothing about them.
When we check attribute presence in the entry, we should strip options
before comparing attribute names with the schema.
Related: https://pagure.io/freeipa/issue/8001
Signed-off-by: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
While [1] did open recursion, it also opened widely a security flaw.
This patch intends to close it back, while allowing operators to easily
add their open configuration within Bind9.
In order to allow operators to still open Bind recursion, a new file is
introduced, "ipa-ext.conf" (path might change according to the OS). This
file is not managed by the installer, meaning changes to it won't be
overridden.
Since it's included at the very end of the main configuration file, it
also allows to override some defaults - of course, operators have to be
careful with that.
Related-Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1754530
Fixes: https://pagure.io/freeipa/issue/8079
[1] 5f4c75eb28
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Stanislav Levin <slev@altlinux.org>
"ipa-advise config-client-for-smart-card-auth" is run on a server and
creates a script that needs to be copied and executed on a client.
The client may be of a different version and use authconfig instead of
authselect. The generated script must be able to handle both cases
(client using authselect or client using authconfig).
The patch checks whether authselect is available and calls the proper
configuration command (authselect or authconfig) depending on its
availability on the client.
Fixes: https://pagure.io/freeipa/issue/8113
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
New Pylint (2.4.3) catches several new 'true problems'. At the same
time, it warns about things that are massively and reasonably
employed in FreeIPA.
list of fixed:
- no-else-continue
- redeclared-assigned-name
- no-else-break
- unnecessary-comprehension
- using-constant-test (false positive)
list of ignored (responsibility of contributors and reviewers):
- import-outside-toplevel
Fixes: https://pagure.io/freeipa/issue/8102
Signed-off-by: Stanislav Levin <slev@altlinux.org>
Reviewed-By: Fraser Tweedale <ftweedal@redhat.com>
Python2 and python3 have different interfaces for os.mkdir:
python2: os.mkdir(path[, mode])
python3: os.mkdir(path, mode=0o777, *, dir_fd=None)
ipa-backup is using the python3 format, which breaks deployments using
python2. The fix consists in using os.mkdir(path, 0o700) instead of
os.mkdir(path, mode=0o700).
Fixes: https://pagure.io/freeipa/issue/8099
Reviewed-By: Alexander Bokovoy <abbra@users.noreply.github.com>
Add an upgrade script to detect when ipakra people entry has
incorrect 'description' attribute and fix it.
Part of: https://pagure.io/freeipa/issue/8084
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
If IPA CA has custom subject DN (not "CN=Certificate
Authority,{subject_base}"), the uid=ipakra people entry gets an
incorrect 'description' attribute. The issuer DN in the
'description' attribute is based on the aforementioned pattern,
instead of the actual IPA CA subject DN.
Update KRAInstance.configure_instance() to require the CA subject DN
argument. Update ipaserver.install.kra.install() to pass the CA
subject DN.
Fixes: https://pagure.io/freeipa/issue/8084
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
The UPG plugin is used to create a user private group when a new
IPA user is created, with the same name as the user. When this plugin
is enabled, the user creation must ensure that no group exists with
the same name.
When the UPG plugin is disabled, or when the user is created with the
--noprivate option, there is no need to perform this check as the
private group will not get created.
Currently, the --noprivate option correctly skips the test, but a
disabled UPG plugin does not skip the test. The fix ensures that
UPG plugin status is checked.
Fixes: https://pagure.io/freeipa/issue/4972
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
This patche removes 93 pylint deprecation warnings due to invalid escape
sequences (mostly 'invalid escape sequence \d') on unicode strings.
Signed-off-by: Rafael Guterres Jeffman <rjeffman@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
This patche removes 93 pylint deprecation warnings due to invalid escape
sequences (mostly 'invalid escape sequence \d') on unicode strings.
Signed-off-by: Rafael Guterres Jeffman <rjeffman@redhat.com>
Reviewed-By: Christian Heimes <cheimes@redhat.com>
Add support for Custodia ca_wrapped clients to specify the desired
symmetric encryption algorithm for exporting the wrapped signing key
(this mechanism is used for LWCA key replication). If not
specified, we must assume that the client has an older Dogtag
version that can only import keys wrapped with DES-EDE3-CBC
encryption.
The selected algorithm gets passed to the 'nsswrappedcert' handler,
which in turn passes it to the 'pki ca-authority-key-export' command
(which is part of Dogtag).
Client-side changes will occur in a subsequent commit.
Part of: https://pagure.io/freeipa/issue/8020
Reviewed-By: Alexander Bokovoy <abbra@users.noreply.github.com>
To support lightweight CA key replication using AES, while retaining
backwards compatibility with old servers, it is necessary to signal
support for AES. Whereas we currently request a key with the path:
/keys/ca_wrapped/<nickname>
and whereas paths with > 3 components are unsupported, add support
for handlers to signal that they support extra arguments (defaulting
to False), those arguments being conveyed as additional path
components, e.g.:
# 2.16.840.1.101.3.4.1.2 = aes128-cbc
/keys/ca_wrapped/<nickname>/2.16.840.1.101.3.4.1.2
This commit only adds the Custodia support for extra handler
arguments. Work to support LWCA key replication with AES wrapping
will continue in subsequent commits.
Part of: https://pagure.io/freeipa/issue/8020
Reviewed-By: Alexander Bokovoy <abbra@users.noreply.github.com>
Helper scripts now use api.bootstrap(log=None) to avoid the creation of
log files. Helper scripts are typically executed from daemons which
perform their own logging. The helpers still log to stderr/stdout.
This also gets rid of some SELinux AVCs when the script tries to write
to /root/.ipa/.
Fixes: https://pagure.io/freeipa/issue/8075
Signed-off-by: Christian Heimes <cheimes@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
When the --server option is provided to ipa-replica-install (1-step
install), make sure that the server offers all the required roles
(CA, KRA). If it's not the case, refuse the installation.
Note that the --server option is ignored when promoting from client to
replica (2-step install with ipa-client-install and ipa-replica-install),
meaning that the existing behavior is not changed in this use case:
by default the host specified in default.conf as server is used for
enrollment, but if it does not provide a required role, another host can
be picked for CA or KRA setup.
Fixes: https://pagure.io/freeipa/issue/7566
Signed-off-by: Florence Blanc-Renaud <flo@redhat.com>
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Mohammad Rizwan Yusuf <myusuf@redhat.com>
ipa installer creates /etc/pkcs11/modules/softhsm2.module in order
to disable global p11-kit configuration for NSS.
This file was not included in the backups, and not restored.
The fix adds the file to the list of files to include in a backup.
Fixes: https://pagure.io/freeipa/issue/8073
Reviewed-By: Rob Crittenden <rcritten@redhat.com>
It was executed after restore_all_files() so PKCS11_MODULES was
already restored so that part was a no-op, but the redhat
restore_pkcs11_modules() also calls unlink() on each restored
file so basically the file would be restored, unlinked, then
since it was already restored, skipped.
By moving the call to restore_pkcs11_modules() earlier it can
do the expected restoration properly.
https://pagure.io/freeipa/issue/8034
Signed-off-by: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
The host password was defined as a Str type so would be
logged in cleartext in the Apache log.
A new class, HostPassword, was defined to only override
safe_value() so it always returns an obfuscated value.
The Password class cannot be used because it has special treatment
in the frontend to manage prompting and specifically doesn't
allow a value to be passed into it. This breaks backwards
compatibility with older clients. Since this class is derived
from Str old clients treat it as a plain string value.
This also removes the search option from passwords.
https://pagure.io/freeipa/issue/8017
Signed-off-by: Rob Crittenden <rcritten@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
If trusted domain object (TDO) is lacking ipaAllowedToPerform;read_keys
attribute values, it cannot be used by SSSD to retrieve TDO keys and the
whole communication with Active Directory domain controllers will not be
possible.
This seems to affect trusts which were created before
ipaAllowedToPerform;read_keys permission granting was introduced
(FreeIPA 4.2). Add back the default setting for the permissions which
grants access to trust agents and trust admins.
Resolves: https://pagure.io/freeipa/issue/8067
Signed-off-by: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
It looks like for some cases we do not have proper set up keytab
retrieval configuration in the old trusted domain object. This mostly
affects two-way trust cases. In such cases, create default configuration
as ipasam would have created when trust was established.
Resolves: https://pagure.io/freeipa/issue/8067
Signed-off-by: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Florence Blanc-Renaud <frenaud@redhat.com>
As new authentication indicators implemented, we also modified server
API to support those new values. Also, "krbprincipalauthind" attribute
is modified to use a pre-defined set of values instead of arbitrary
strings.
Resolves: https://pagure.io/freeipa/issue/8001
Signed-off-by: Changmin Teng <cteng@redhat.com>
Reviewed-By: Alexander Bokovoy <abokovoy@redhat.com>
Reviewed-By: Simo Sorce <ssorce@redhat.com>
Reviewed-By: Robbie Harwood <rharwood@redhat.com>
When parsing a keytab to copy keys to a different keytab, we don't need
the timestamp, so don't ask klist to output it. In some locales (en_IN,
for example), the timestamp is output in a single field without a space
between date and time. In other locales it can be represented with date
and time separated by a space.
Fixes: https://pagure.io/freeipa/issue/8066
Reviewed-By: Thomas Woerner <twoerner@redhat.com>
There is a loop which keeps trying to bind as the admin user
which will fail until it is replicated.
In the case where there is a lot to replicate the default
5 minute timeout may be insufficient. Provide a hint for
tuning.
Fixes: https://pagure.io/freeipa/issue/7971
Reviewed-By: Florence Blanc-Renaud <flo@redhat.com>
