| Summary: | IPA throws attribute error when trying to add a new default objectclass | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 6 | Reporter: | Pete Savage <psavage> |
| Component: | ipa | Assignee: | Martin Kosek <mkosek> |
| Status: | CLOSED INSUFFICIENT_DATA | QA Contact: | Namita Soman <nsoman> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | unspecified | ||
| Version: | 6.4 | CC: | abokovoy, psavage, rcritten |
| Target Milestone: | rc | ||
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2013-11-04 10:59:36 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Pete Savage
2013-10-18 15:21:55 UTC
I've tracked this down to the fact that ipaKrbAuthzData attribute is not in any of object classes we use for cn=ipaConfig,cn=etc,$SUFFIX. Since the entry is created in the bootstrap template, it is not validated. However, config-mod --userobjectclasses does some operations that cause validation of the schema correctness for the entry and that one fails. The fix would be to add ipaKrbAuthzData to an object class that could be added to cn=ipaConfig,cn=etc,$SUFFIX, and also write an update plugin that fixes the entry on package updates. The attributeType is already there, in ipaGuiConfig objectClass. It is being added in install/updates/10-selinuxusermap.update. Pete, can you please run the following ldapsearch so that we can see why the attribute is not allowed? $ ldapsearch -h $IPA_FQDN -D "cn=Directory Manager" -W -x -b "cn=schema" objectClasses | grep -A 8 -i ipaguiconfig On a side note, did you upgrade IPA from some very old version? Or was it a new installation? Hi Martin, SASL/GSSAPI authentication started SASL username: admin@$IPA_FQDN SASL SSF: 56 SASL data security layer installed. objectClasses: ( 2.16.840.1.113730.3.8.2.1 NAME 'ipaGuiConfig' AUXILIARY MAY ( ipaUserSearchFields $ ipaGroupSearchFields $ ipaSearchTimeLimit $ ipaSearchR ecordsLimit $ ipaCustomFields $ ipaHomesRootDir $ ipaDefaultLoginShell $ ipaD efaultPrimaryGroup $ ipaMaxUsernameLength $ ipaPwdExpAdvNotify $ ipaUserObjec tClasses $ ipaGroupObjectClasses $ ipaDefaultEmailDomain $ ipaMigrationEnable d $ ipaCertificateSubjectBase $ ipaSELinuxUserMapDefault $ ipaSELinuxUserMapO rder ) ) objectClasses: ( 2.16.840.1.113730.3.8.4.13 NAME 'ipaConfigObject' DESC 'gener ic config object for IPA' AUXILIARY MAY ipaConfigString X-ORIGIN 'IPA v2' ) This was a fresh install. That's strange, LDAP update procedure done as a part of IPA installation should have updated the objectClass as is. Were there any errors in ipa-replica-install in that stage? This issue may be fixed by simply running the LDAP Update step again, with this command # ipa-ldap-updater --upgrade run on the IPA server. I worked with Pete on this Bugzilla. It seemed as an installation issue where LDAP Update was probably not run during an installation as ipaGuiConfig objectclass was not extended by install/updates/10-selinuxusermap.update. Pete, please provide ipaserver-install.log so that we can double check that. Without the log, it is hard to guess what might have gone wrong. If you do not have the log ready, we would have to close the bug. Pete provided ipaserver-install.log to me privately. I investigated it and the ipaGuiConfig objectClass was updated, according to the log. Additionally, ipakrbauthz attribute was added to the config object which could not have worked if the attribute was not allowed: 2013-10-17T11:39:08Z DEBUG Final value after applying updates 2013-10-17T11:39:08Z DEBUG dn: cn=ipaConfig,cn=etc,dc=example,dc=com 2013-10-17T11:39:08Z DEBUG ipadefaultloginshell: /bin/sh 2013-10-17T11:39:08Z DEBUG ipacertificatesubjectbase: O=EXAMPLE.COM 2013-10-17T11:39:08Z DEBUG ipamigrationenabled: FALSE 2013-10-17T11:39:08Z DEBUG cn: ipaConfig 2013-10-17T11:39:08Z DEBUG ipaSELinuxUserMapDefault: 2013-10-17T11:39:08Z DEBUG unconfined_u:s0-s0:c0.c1023 2013-10-17T11:39:08Z DEBUG ipaUserObjectClasses: 2013-10-17T11:39:08Z DEBUG top 2013-10-17T11:39:08Z DEBUG person 2013-10-17T11:39:08Z DEBUG organizationalperson 2013-10-17T11:39:08Z DEBUG inetorgperson 2013-10-17T11:39:08Z DEBUG inetuser 2013-10-17T11:39:08Z DEBUG posixaccount 2013-10-17T11:39:08Z DEBUG krbprincipalaux 2013-10-17T11:39:08Z DEBUG krbticketpolicyaux 2013-10-17T11:39:08Z DEBUG ipaobject 2013-10-17T11:39:08Z DEBUG ipasshuser 2013-10-17T11:39:08Z DEBUG ipadefaultprimarygroup: ipausers 2013-10-17T11:39:08Z DEBUG ipapwdexpadvnotify: 4 2013-10-17T11:39:08Z DEBUG objectclass: 2013-10-17T11:39:08Z DEBUG nsContainer 2013-10-17T11:39:08Z DEBUG top 2013-10-17T11:39:08Z DEBUG ipaGuiConfig 2013-10-17T11:39:08Z DEBUG ipaConfigObject 2013-10-17T11:39:08Z DEBUG ipagroupsearchfields: cn,description 2013-10-17T11:39:08Z DEBUG ipadefaultemaildomain: example.com 2013-10-17T11:39:08Z DEBUG ipahomesrootdir: /home 2013-10-17T11:39:08Z DEBUG ipasearchtimelimit: 2 2013-10-17T11:39:08Z DEBUG ipaKrbAuthzData: 2013-10-17T11:39:08Z DEBUG MS-PAC 2013-10-17T11:39:08Z DEBUG ipagroupobjectclasses: 2013-10-17T11:39:08Z DEBUG top 2013-10-17T11:39:08Z DEBUG groupofnames 2013-10-17T11:39:08Z DEBUG nestedgroup 2013-10-17T11:39:08Z DEBUG ipausergroup 2013-10-17T11:39:08Z DEBUG ipaobject 2013-10-17T11:39:08Z DEBUG ipasearchrecordslimit: 100 2013-10-17T11:39:08Z DEBUG ipaSELinuxUserMapOrder: 2013-10-17T11:39:08Z DEBUG guest_u:s0$xguest_u:s0$user_u:s0$staff_u:s0-s0:c0.c1023$unconfined_u:s0-s0:c0.c1023 2013-10-17T11:39:08Z DEBUG ipaconfigstring: AllowNThash 2013-10-17T11:39:08Z DEBUG ipamaxusernamelength: 32 2013-10-17T11:39:08Z DEBUG ipausersearchfields: uid,givenname,sn,telephonenumber,ou,title 2013-10-17T11:39:08Z DEBUG [(0, 'ipakrbauthzdata', ['MS-PAC'])] 2013-10-17T11:39:08Z DEBUG Live 1, updated 1 2013-10-17T11:39:08Z INFO Done I am thinking that the objectClass would have to been replaced post-install, by some other installation. Maybe a really old replica which would replace the objectClass definition without looking at current value? Anyway, I am leaving needsinfo? on Pete to try to provide some reproduction scenario which could help us identify what could possibly went wrong. Without reproduction scenario, I would need to close this bug - we can reopen when/if the issue occurs again and we would have a hint about the scenario. We do not have any reproduction scenario or information that could lead us to the potential issue. I will therefore close this bug. However, please feel free to reach back and re-open if you find information how to reproduce. |