URL: https://github.com/SSSD/sssd/pull/127
Author: pbrezina
Title: #127: ssh: use cache_req
Action: opened
PR body:
"""
This patches makes SSH responder use the cache_req interface.
Since this responder uses that same `cache-domain-cache` lookup logic
for host certificates I implemented `host by name` request in `cache_req`.
In order to achieve this I moved data provider lookup function from `cache_req`
code into plugins.
The first two patches fixes minor issues in the SSH responder and should be
pushed to earlier versions as well. The first patch fix a little issue
introduced probably by overrides and the second patch removes name qualification
since it is already qualified in the sysdb since fqname patches.
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/127/head:pr127
git checkout pr127
URL: https://github.com/SSSD/sssd/pull/141
Author: fidencio
Title: #141: PAM: Use cache_req_user_by_name_*()
Action: opened
PR body:
"""
PAM responder has been already using a lot from cache_req and one of the
missing parts (most likely the only one?) was that
pam_check_user_search() has been using sss_dp_get_account_send(), which
could be dropped (and a lot of code around it) in favour of using
cache_req_user_by_name_*() instead.
Resolves:
https://fedorahosted.org/sssd/ticket/1126
Signed-off-by: Fabiano FidĂȘncio <fidencio(a)redhat.com>
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/141/head:pr141
git checkout pr141
URL: https://github.com/SSSD/sssd/pull/137
Author: sumit-bose
Title: #137: Initial pkinit support
Action: opened
PR body:
"""
This series of patches add initial support for PKINIT
(https://fedorahosted.org/sssd/ticket/3270) by forwarding the information about
the selected certificate from the Smartcard to the backends. Currently only
the krb5 backend supports Smartcard authentication the other backends will
return an error code which will tell the PAM responder to fall back to local
Smartcard authentication as it is currently the case.
Testing requires a working PKINIT setup which e.g. can be done with AD by
setting up a CA and generating certificates as described in
https://fedorahosted.org/sssd/wiki/DesignDocs/SmartcardAuthenticationTestinâŠ.
But currently more important is regression testing, i.e. making sure all other
authentication methods are still working as expected.
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/137/head:pr137
git checkout pr137
URL: https://github.com/SSSD/sssd/pull/70
Author: sumit-bose
Title: #70: check_duplicate: check name member before using it
Action: opened
PR body:
"""
The second patch resolves https://fedorahosted.org/sssd/ticket/3231 and adds a
test for the issue.
The first patches fixes a potential memory leak which so far was only relevant
in the tests.
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/70/head:pr70
git checkout pr70
URL: https://github.com/SSSD/sssd/pull/66
Author: justin-stephenson
Title: #66: Minor Dynamic DNS fixes
Action: opened
PR body:
"""
To provide a bit more information, one of the fixes is to correct NULL being printed here(https://fedorahosted.org/sssd/ticket/3220):
[nsupdate_msg_create_common] (0x0200): Creating update message for realm [(null)].
For the other(https://bugzilla.redhat.com/show_bug.cgi?id=1386748), It is not uncommon for nsupdate to successfully update DNS records but report the error below which results in return(2) to be called inside nsupdate code
TSIG error with server: tsig verify failure
It is easy to reproduce with AD DNS changing Dynamic DNS to 'Nonsecure and secure' on the Zone Properties.
This patch allows PTR records to continue when this happens, however in this case our debug log messages still report failure and I think some improvement should be made here(not sure how exactly though)
[child_sig_handler] (0x1000): Waiting for child [3710].
[nsupdate_child_handler] (0x0040): Dynamic DNS child failed with status [512]
[child_sig_handler] (0x0020): child [3710] failed with status [2].
[be_nsupdate_done] (0x0040): nsupdate child execution failed [1432158238]: Dynamic DNS update failed
It would be nice to correct this at the nsupdate level if this is not the expected behavior also.
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/66/head:pr66
git checkout pr66
URL: https://github.com/SSSD/sssd/pull/106
Author: jhrozek
Title: #106: Add a new "files" provider
Action: opened
PR body:
"""
This patch set implements a new provider that mirrors the contents of passwd and groups files. The intent is to make these users and groups available through the SSSD memory cache to improve performance and make it possible to read extended attributes via the sssd D-Bus interface.
As the next step, we will implement a writable D-Bus interface to make it possible to also manage local users with an API and supersede the AccountService API.
At the moment, the domain must be enabled explicitly. When this branch is merged, another commit would also enable the files domain for all installations by default.
In order to make the resolution precise, the files domain is disabled once an inotify notification arrives During testing, I realized there might be a delay between changing the UNIX files by replacing them and *receiving* the inotify notification. Therefore the tests add a sleep as well. That's not nice and I would be glad if the reviewer can spot how to speed up the inotify notification receiving.
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/106/head:pr106
git checkout pr106
URL: https://github.com/SSSD/sssd/pull/85
Author: celestian
Title: #85: SYSDB: Removing of sysdb_try_to_find_expected_dn()
Action: opened
PR body:
"""
Currently in order to match multiple LDAP search results we
use two different functions - we have sysdb_try_to_find_expected_dn()
but also sdap_object_in_domain().
This patch removes sysdb_try_to_find_expected_dn() and add new
sdap_search_initgr_user_in_batch() based on sdap_object_in_domain().
This function covers necessary logic.
Resolves:
https://fedorahosted.org/sssd/ticket/3230
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/85/head:pr85
git checkout pr85
URL: https://github.com/SSSD/sssd/pull/138
Author: justin-stephenson
Title: #138: IPA: Skip conflict entries associated with sudo rules
Action: opened
PR body:
"""
SSSD retrieves sudo rule information from the IPA LDAP tree, conflict entries will cause problems for SSSD and disallow sudo access when SSSD code is parsing entries associated with sudo rules. This PR sets a skip_entry boolean when it is appropriate and skips over these conflict entries.
Ticket: https://fedorahosted.org/sssd/ticket/3288
Reproducer steps: Create host conflict entry and associate it with a sudo rule that is assigned to certain hosts, attempt to sudo as IDM user. I had some difficulty attempting to force replication issues causing the creation of a conflict entry, the below manual ldapmodify steps will work also:
- Retrieve the DN of the sudoRule
`# ipa sudorule-find --all --raw | grep 'dn: '
dn: ipaUniqueID=e9025c46-ddab-11e6-9096-525400af7498,cn=sudorules,cn=sudo,dc=jstephen,dc=local`
- Run ldapmodify similar to below
dn: ipaUniqueID=e9025c46-ddab-11e6-9096-525400af7498,cn=sudorules,cn=sudo,dc=jstephen,dc=local
changetype: modify
add: memberHost
memberHost: fqdn=testhost.jstephen.local+nsuniqueid=cb3d7383-ddb511e6-8c9996c1-71a1e36a,cn=computers,cn=accounts,dc=jstephen,dc=local
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/138/head:pr138
git checkout pr138
URL: https://github.com/SSSD/sssd/pull/108
Author: lslebodn
Title: #108: Partially revert "CONFIG: Use default config when none provided"
Action: opened
PR body:
"""
This reverts part of commit 59744cff6edb106ae799b2321cb8731edadf409a.
Removed is copying of default configuration into /etc/sssd/sssd.conf
Sample configurations is still part of installation.
Copying default configuration from /usr/lib64/sssd/conf/sssd.conf -> /etc/sssd/sssd.conf
is not the best idea. There are better way how to use default configuration
and we will need to change anyway due to files provider.
I can revert sample configuration as well.
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/108/head:pr108
git checkout pr108
URL: https://github.com/SSSD/sssd/pull/112
Author: justin-stephenson
Title: #112: FAILOVER: Improve port status log messages
Action: opened
PR body:
"""
This PR is created to ensure administrators do not follow the wrong path troubleshooting issues, especially with messages such as:
`Port status of port 636 for server 'server' is 'not working'`
This patch should help to clarify that there is no direct relationship between port status and actual networking status.
"""
To pull the PR as Git branch:
git remote add ghsssd https://github.com/SSSD/sssd
git fetch ghsssd pull/112/head:pr112
git checkout pr112