In summary this patch does:
* Follow the defined rules for cookies when:
- receiving a cookie (process the attributes)
- storing a cookie (store cookie + attributes)
- sending a cookie
+ validate the cookie domain against the request URL
+ validate the cookie path against the request URL
+ validate the cookie expiration
+ if valid then send only the cookie, no attribtues
* Modifies how a request URL is stored during a XMLRPC
request/response sequence.
* Refactors a bit of the request/response logic to allow for making
the decision whether to send a session cookie instead of full
Kerberous auth easier.
* The server now includes expiration information in the session cookie
it sends to the client. The server always had the information
available to prevent using an expired session cookie. Now that
expiration timestamp is returned to the client as well and now the
client will not send an expired session cookie back to the server.
* Adds a new module and unit test for cookies (see below)
Formerly we were always returning the session cookie no matter what
the domain or path was in the URL. We were also sending the cookie
attributes which are for the client only (used to determine if to
return a cookie). The attributes are not meant to be sent to the
server and the previous behavior was a protocol violation. We also
were not checking the cookie expiration.
Cookie library issues:
We need a library to create, parse, manipulate and format cookies both
in a client context and a server context. Core Python has two cookie
libraries, Cookie.py and cookielib.py. Why did we add a new cookie
module instead of using either of these two core Python libaries?
Cookie.py is designed for server side generation but can be used to
parse cookies on the client. It's the library we were using in the
server. However when I tried to use it in the client I discovered it
has some serious bugs. There are 7 defined cookie elements, it fails
to correctly parse 3 of the 7 elements which makes it unusable because
we depend on those elements. Since Cookie.py was designed for server
side cookie processing it's not hard to understand how fails to
correctly parse a cookie because that's a client side need. (Cookie.py
also has an awkward baroque API and is missing some useful
functionality we would have to build on top of it).
cookielib.py is designed for client side. It's fully featured and obeys
all the RFC's. It would be great to use however it's tightly coupled
with another core library, urllib2.py. The http request and response
objects must be urllib2 objects. But we don't use urllib2, rather we use
httplib because xmlrpclib uses httplib. I don't see a reason why a
cookie library should be so tightly coupled to a protocol library, but
it is and that means we can't use it (I tried to just pick some isolated
entrypoints for our use but I kept hitting interaction/dependency problems).
I decided to solve the cookie library problems by writing a minimal
cookie library that does what we need and no more than that. It is a
new module in ipapython shared by both client and server and comes
with a new unit test. The module has plenty of documentation, no need
to repeat it here.
Request URL issues:
We also had problems in rpc.py whereby information from the request
which is needed when we process the response is not available. Most
important was the requesting URL. It turns out that the way the class
and object relationships are structured it's impossible to get this
information. Someone else must have run into the same issue because
there was a routine called reconstruct_url() which attempted to
recreate the request URL from other available
information. Unfortunately reconstruct_url() was not callable from
inside the response handler. So I decided to store the information in
the thread context and when the request is received extract it from
the thread context. It's perhaps not an ideal solution but we do
similar things elsewhere so at least it's consistent. I removed the
reconstruct_url() function because the exact information is now in the
context and trying to apply heuristics to recreate the url is probably
not robust.
Ticket https://fedorahosted.org/freeipa/ticket/3022
The configuration code has been modified to use the ConfigParser to
set the parameters in the CA section in the deployment configuration.
This allows IPA to define additional PKI subsystems in the same
configuration file.
PKI Ticket #399 (https://fedorahosted.org/pki/ticket/399)
OpenSSH server included in Fedora 18 raises a validation error when
the tested AuthorizedKeysCommand/PubKeyAgent option is tested with
an empty value. It requires a command with an absolute path to be
passed. Due to this issue, sshd support is never configured on
Fedora 18.
Pass the real agent we will use later to the testing command to
avoid this error.
Windows 2012 Server changed procedure how KERB_VALIDATION_INFO ([MS-PAC]
section 2.5) is populated. Detailed description is available in [MS-KILE]
version 25.0 and above.
Refactor KERB_VALIDATION_INFO verification and ensure we filter out extra
SIDs in case they belong to our domain.
https://fedorahosted.org/freeipa/ticket/3231
Fedora 16 introduced chrony as default client time&date synchronization
service:
http://fedoraproject.org/wiki/Features/ChronyDefaultNTP
Thus, there may be people already using chrony as their time and date
synchronization service before installing IPA.
However, installing IPA server or client on such machine may lead to
unexpected behavior, as the IPA installer would configure ntpd and leave
the machine with both ntpd and chronyd enabled. However, since the OS
does not allow both chronyd and ntpd to be running concurrently and chronyd
has the precedence, ntpd would not be run on that system at all.
Make sure, that user is warned when trying to install IPA on such
system and is given a possibility to either not to let IPA configure
ntpd at all or to let the installer stop and disable chronyd.
https://fedorahosted.org/freeipa/ticket/2974
Modify the default IPA CA certificate profile to include CRL and
OCSP extensions which will add URIs to IPA CRL&OCSP to published
certificates.
Both CRL and OCSP extensions have 2 URIs, one pointing directly to
the IPA CA which published the certificate and one to a new CNAME
ipa-ca.$DOMAIN which was introduced as a general CNAME pointing
to all IPA replicas which have CA configured.
The new CNAME is added either during new IPA server/replica/CA
installation or during upgrade.
https://fedorahosted.org/freeipa/ticket/3074https://fedorahosted.org/freeipa/ticket/1431
There were a number of code paths where we would try to call
xmlrpc_env_clean() without having first called xmlrpc_env_init()
Re-order the code so we always initialize the XML-RPC client first.
I also noticed a place where the return value of strdup() was not
being checked for NULL.
https://fedorahosted.org/freeipa/ticket/3275
Wrap the password change extop in a transaction.
Fix the case where a password is reset and then immediately used. If done
fast enough then the KDC may not detect that the password is expired and
grant access using the expired password rather than prompting for a reset.
https://fedorahosted.org/freeipa/ticket/1064
When user from other realm than FreeIPA's tries to use Web UI (login via forms-based auth or with valid trusted realm ticket), he gets an unauthorized error with X-Ipa-Rejection-Reason=denied. Web UI responds with showing login dialog with following error message: 'Sorry you
are not allowed to access this service.'.
Note: such users are not supported because they don't have a corresponding entry in LDAP which is needed for ACLs.
https://fedorahosted.org/freeipa/ticket/3252
denied change
When user from other realm than FreeIPA's tries to use Web UI
(login via forms-based auth or with valid trusted realm ticket),
the 401 Unauthorized error with X-Ipa-Rejection-Reason=denied
is returned.
Also, the support for usernames of the form user@SERVER.REALM
or user@server.realm was added.
https://fedorahosted.org/freeipa/ticket/3252
When DNS zone/record manipulation commands fails for example due to
a ValidationError, ipa-replica-prepapre reports a whole traceback
which is difficult to read. Make sure our error error is more
readable.
https://fedorahosted.org/freeipa/ticket/3283
Ther3 are two global ipaConfig options to disable undesirable writes that
have performance impact.
The "KDC:Disable Last Success" will disable writing back to ldap the last
successful AS Request time (successful kinit)
The "KDC:Disable Lockout" will disable completely writing back lockout
related data. This means lockout policies will stop working.
https://fedorahosted.org/freeipa/ticket/2734
The best way to configure DNS depends on the environment and no general
recommendations should be given by the CLI or Web UI. Especially
forwarders should not be recommended by only be option of last resort.
Fixes https://fedorahosted.org/freeipa/ticket/3261
Recent versions of authconfig do not restart sssd if only the
--enablesssd and --enablesssdauth options are used. To make sure sssd is
running after ipa-server-install is run this patch add an unconditional
restart of sssd after authconfig is run during the installation.
Since there already is some logic trying to determine if sssd needs to
be restarted or stopped if freeipa in uninstalled no changes are needed
here.
Fixes https://fedorahosted.org/freeipa/ticket/3267
Currently only the group SIDs from a PAC are used to find out about the
membership in local groups. This patch adds the user SID to the list.
Fixes https://fedorahosted.org/freeipa/ticket/3257
The current Linux NFS server is severely limited when it comes to handling
kerberos tickets. Bsically any ticket bigger than 2k will cause it to fail
authentication due to kernel->userspace upcall interface restrictions.
Until we have additional support in IPA to indivdually mark principals to
opt out of getting PACs attached we always prevent PACs from being attached
to TGTs or Tickets where NFS is involved.
Incorporate SELinux policy changes introduced in Dogtag 10 in IPA
SELinux policy:
- dogtag10 now runs with pki_tomcat_t context instead of pki_ca_t
- certmonger related rule are now integrated in system policy and
can be removed from IPA policy
Also remove redundant SELinux rules for connection of httpd_t, krb5kdc_t
or named_t to DS socket. The socket has different target type anyway
(dirsrv_var_run_t) and the policy allowing this is already in
system.
https://fedorahosted.org/freeipa/ticket/3234
After upgrade, sshkeys of existing users and hosts or mac address are not editable because attribute level rights are not send to Web UI due to lack of ipasshuser/ieee802device object classes.
'w_if_no_aci' attribute flag was introduced to bypass this issue. It makes attribute writable when AttributeLevelRights for the attribute are not present and only when user posses rights for modifying object class attribute.
The flag was set for sshkeys_field and mac address field.
https://fedorahosted.org/freeipa/ticket/3260
In FreeIPA 2.2, the default for group type was 'POSIX', in 3.0 it's 'normal' (non-posix). CLI's default is 'POSIX', Web UI should use 'POSIX' as well.
https://fedorahosted.org/freeipa/ticket/3270
When using Global Catalog for resolving users and groups, one needs
to authenticate against Active Directory DC using GSSAPI. Obtaining
Kerberos ticket might fail. Make sure the failure message is actually
returned to the user so that decision can be made about the source
of issue.
https://fedorahosted.org/freeipa/ticket/3265
The ipa-csreplica-manage tool often assumed that the port numbers are the
same on both sides of a replication agreement.
This assumption doesn't hold in a cluster with both old-style hosts and
ones with merged DBs.
When managing agreements, determine the port with the PKI (or merged) DS
on each master, and use it.
Also, in CSReplicationManager, always use starttls rather than ldaps://.
The ancient IPAdmin class used some heuristics to determine the protocol
to connect with (ldap, ldaps, or ldapi). In turn, some calling code used
questionable mechanisms to get the correct protocol: in
ipaserver/install/replication.py, the CA cert was either passed to the
constructor or added to the class afterwards, to get ldap:// or ldaps://.
Add an explicit protocol argument and only fall back to backwards-compatible
guessing if it is not given.
With the new unified Dogtag10 LDAP database, PKI-CA data and the
agreements themselves are now in the main LDAP instance.
Replication management tools now need to properly filter replication
agreements based on the suffix to avoid clashing of agreements of
different types.
Stopping certificate tracking was done as part of the PKI DS uninstall.
Since with the merged DB, thePKI DS is not used any more, this step
was skipped.
Move certificate untracking to a separate step and call it separately.
Also, the post-uninstall check for tracked certificates used the wrong
set of Dogtag constants. Fix the issue.
This brings /etc/dirsrv/slapd-REALM/certmap.conf under IPA control.
The file is overwritten on upgrades.
This ensures that the cert for the ipaca user is recognized when
ipa-ca-install is run on older masters.
The new merged database will replicate with both the IPA and CA trees, so all
DS instances (IPA and CA on the existing master, and the merged one on the
replica) need to have the same schema.
Dogtag does all its schema modifications online. Those are replicated normally.
The basic IPA schema, however, is delivered in ldif files, which are not
replicated. The files are not present on old CA DS instances. Any schema
update that references objects in these files will fail.
The whole 99user.ldif (i.e. changes introduced dynamically over LDAP) is
replicated as a blob. If we updated the old master's CA schema dynamically
during replica install, it would conflict with updates done during the
installation: the one with the lower CSN would get lost.
Dogtag's spawn script recently grew a new flag, 'pki_clone_replicate_schema'.
Turning it off tells Dogtag to create its schema in the clone, where the IPA
modifications are taking place, so that it is not overwritten by the IPA schema
on replication.
The patch solves the problems by:
- In __spawn_instance, turning off the pki_clone_replicate_schema flag.
- Providing a script to copy the IPA schema files to the CA DS instance.
The script needs to be copied to old masters and run there.
- At replica CA install, checking if the schema is updated, and failing if not.
The --skip-schema-check option is added to ipa-{replica,ca}-install to
override the check.
All pre-3.1 CA servers in a domain will have to have the script run on them to
avoid schema replication errors.
https://fedorahosted.org/freeipa/ticket/3213
New servers that are installed with dogtag 10 instances will use
a single database instance for dogtag and IPA, albeit with different
suffixes. Dogtag will communicate with the instance through a
database user with permissions to modify the dogtag suffix only.
This user will authenticate using client auth using the subsystem cert
for the instance.
This patch includes changes to allow the creation of masters and clones
with single ds instances.
python-crypto package is not available everywhere, use m2crypto instead.
Originally we thought to extend python-krbV to provide krb5_c_encrypt()
wrapper but m2crypto is readily available.
https://fedorahosted.org/freeipa/ticket/3271
FreeIPA 3.0 is being released to Fedora 18 only. Since we only support
Fedora 17 and Fedora 18 in FreeIPA 3.0+, compatibility code for older
Fedoras can be dropped. This should clean up the spec file and make it
more readable.
Dogtag10 Requires were fixed. Without this patch, there is a conflict
on dogtag-pki-common-theme.
Tar requirement was added to avoid crashes in ipa-replica-prepare on
some minimal Fedora composes.
https://fedorahosted.org/freeipa/ticket/2748https://fedorahosted.org/freeipa/ticket/3237
Web UI part of 'disable global forwaring per zone' effort.
Option "Forwarding disabled" was added to 'DNS global config' and 'DNS zone' forwarding policy. It corresponds to 'none' value of idnsforwardpolicy.
https://fedorahosted.org/freeipa/ticket/3209
bind-dyndb-ldap allows disabling global forwarder per-zone. This may
be useful in a scenario when we do not want requests to delegated
sub-zones (like sub.example.com. in zone example.com.) to be routed
through global forwarder.
Few lines to help added to explain the feature to users too.
https://fedorahosted.org/freeipa/ticket/3209
As named.conf and bind-dyndb-plugin is not set up yet during DNS
configuration phase, IPA hostname (i.e. the nameserver) should not
be required be to resolvable in this phase.
https://fedorahosted.org/freeipa/ticket/3248
