Re: [389-users] Stumped - SSL works for auth, sudo, etc, but fails for ldap user cronjobs
Hi,
what kind of certificate do you use, selfsigned? Are the certificates signed by the same CA?
Am 18.07.12, schrieb David Nguyen d_k_nguyen@yahoo.com:
Hi all,
I have a strange one. My current setup is working perfectly. client1 is able to connect to ldap-server1 via SSL and everything is working correctly. I then had a need to add another ldap server (ldap-server2) as a multi-master replica and everything is working (user auth, sudo via ldap users, ldapsearch, openssl, etc) except cronjobs for users served out of ldap fail to run.
I can see this in the error log on ldap-server2:
[18/Jul/2012:11:18:00 -0700] - PR_Recv for connection 467 returns -12195 (Peer does not recognize and trust the CA that issued your certificate.)
If I set /etc/ldap.conf to not use SSL (URI ldap://fqdn vs URI ldaps://fqdn:636), the cronjobs fire just fine.
So it appears as though there is an SSL cert issue, but I'm stumped because all of the other services that use ldap on client1 work except cron jobs (root cron fires fine as expected since nsswitch is set to files then ldap).
If I replace the URI string in /etc/ldap.conf to point at ldap-server1, cron starts working.
Both ldap-server1 and ldap-server2 are using running the same OS and kernel version (RHEL5) as well as the same version of 389 DS (389-ds-1.2.1-1.el5).
Any ideas as to what could be causing this problem? Here is the /etc/ldap.conf on client1 if it matters:
====== begin /etc/ldap.conf ======= URI ldaps://ops-ldap006.svale.netledger.com:636 base dc=netsuite,dc=com
timelimit 10 bind_policy soft nss_reconnect_tries 3 bind_timelimit 6 idle_timelimit 30 sudoers_base ou=SUDOers,dc=netsuite,dc=com sudoers_debug 0
##ssl start_tls TLS_CACERT /etc/openldap/cacerts/ca.crt TLS_CACERTFILE /etc/openldap/cacerts/ca.crt TLS_REQCERT demand pam_lookup_policy yes pam_password exop
nss_initgroups_ignoreusers root,named,avahi,haldaemon,dbus,gdm,postfix,puppet
====== end /etc/ldap.conf =======
Thanks in advance, David -- 389 users mailing list 389-users@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/389-users
-- Carsten Grzemba
The cert is self-signed, but by different CA's (each server has it's own CA).
You know what? I took your hint and signed a new server cert using the "working" ldap server's CA and voila, it started working. Thank you so much! I've been scratching my head over this one for days
David
On Thu, Jul 19, 2012 at 6:34 AM, Carsten Grzemba grzemba@contac-dt.de wrote:
Hi,
what kind of certificate do you use, selfsigned? Are the certificates signed by the same CA?
Am 18.07.12, schrieb David Nguyen d_k_nguyen@yahoo.com:
Hi all,
I have a strange one. My current setup is working perfectly. client1 is able to connect to ldap-server1 via SSL and everything is working correctly. I then had a need to add another ldap server (ldap-server2) as a multi-master replica and everything is working (user auth, sudo via ldap users, ldapsearch, openssl, etc) except cronjobs for users served out of ldap fail to run.
I can see this in the error log on ldap-server2:
[18/Jul/2012:11:18:00 -0700] - PR_Recv for connection 467 returns -12195 (Peer does not recognize and trust the CA that issued your certificate.)
If I set /etc/ldap.conf to not use SSL (URI ldap://fqdn vs URI ldaps://fqdn:636), the cronjobs fire just fine.
So it appears as though there is an SSL cert issue, but I'm stumped because all of the other services that use ldap on client1 work except cron jobs (root cron fires fine as expected since nsswitch is set to files then ldap).
If I replace the URI string in /etc/ldap.conf to point at ldap-server1, cron starts working.
Both ldap-server1 and ldap-server2 are using running the same OS and kernel version (RHEL5) as well as the same version of 389 DS (389-ds-1.2.1-1.el5).
Any ideas as to what could be causing this problem? Here is the /etc/ldap.conf on client1 if it matters:
====== begin /etc/ldap.conf ======= URI ldaps://ops-ldap006.svale.netledger.com:636 base dc=netsuite,dc=com
timelimit 10 bind_policy soft nss_reconnect_tries 3 bind_timelimit 6 idle_timelimit 30 sudoers_base ou=SUDOers,dc=netsuite,dc=com sudoers_debug 0
##ssl start_tls TLS_CACERT /etc/openldap/cacerts/ca.crt TLS_CACERTFILE /etc/openldap/cacerts/ca.crt TLS_REQCERT demand pam_lookup_policy yes pam_password exop
nss_initgroups_ignoreusers root,named,avahi,haldaemon,dbus,gdm,postfix,puppet
====== end /etc/ldap.conf =======
Thanks in advance, David -- 389 users mailing list 389-users@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/389-users
-- Carsten Grzemba
-- 389 users mailing list 389-users@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/389-users
Just as a follow up to this, on ~5% of our hosts (RHEL[456]), crond would be unable to connect to the ldapserver after /etc/ldap.conf was updated to use SSL. Restarting crond fixed the issue.
On Thu, Jul 19, 2012 at 10:54 AM, David Nguyen d_k_nguyen@yahoo.com wrote:
The cert is self-signed, but by different CA's (each server has it's own CA).
You know what? I took your hint and signed a new server cert using the "working" ldap server's CA and voila, it started working. Thank you so much! I've been scratching my head over this one for days
David
On Thu, Jul 19, 2012 at 6:34 AM, Carsten Grzemba grzemba@contac-dt.de wrote:
Hi,
what kind of certificate do you use, selfsigned? Are the certificates signed by the same CA?
Am 18.07.12, schrieb David Nguyen d_k_nguyen@yahoo.com:
Hi all,
I have a strange one. My current setup is working perfectly. client1 is able to connect to ldap-server1 via SSL and everything is working correctly. I then had a need to add another ldap server (ldap-server2) as a multi-master replica and everything is working (user auth, sudo via ldap users, ldapsearch, openssl, etc) except cronjobs for users served out of ldap fail to run.
I can see this in the error log on ldap-server2:
[18/Jul/2012:11:18:00 -0700] - PR_Recv for connection 467 returns -12195 (Peer does not recognize and trust the CA that issued your certificate.)
If I set /etc/ldap.conf to not use SSL (URI ldap://fqdn vs URI ldaps://fqdn:636), the cronjobs fire just fine.
So it appears as though there is an SSL cert issue, but I'm stumped because all of the other services that use ldap on client1 work except cron jobs (root cron fires fine as expected since nsswitch is set to files then ldap).
If I replace the URI string in /etc/ldap.conf to point at ldap-server1, cron starts working.
Both ldap-server1 and ldap-server2 are using running the same OS and kernel version (RHEL5) as well as the same version of 389 DS (389-ds-1.2.1-1.el5).
Any ideas as to what could be causing this problem? Here is the /etc/ldap.conf on client1 if it matters:
====== begin /etc/ldap.conf ======= URI ldaps://ops-ldap006.svale.netledger.com:636 base dc=netsuite,dc=com
timelimit 10 bind_policy soft nss_reconnect_tries 3 bind_timelimit 6 idle_timelimit 30 sudoers_base ou=SUDOers,dc=netsuite,dc=com sudoers_debug 0
##ssl start_tls TLS_CACERT /etc/openldap/cacerts/ca.crt TLS_CACERTFILE /etc/openldap/cacerts/ca.crt TLS_REQCERT demand pam_lookup_policy yes pam_password exop
nss_initgroups_ignoreusers root,named,avahi,haldaemon,dbus,gdm,postfix,puppet
====== end /etc/ldap.conf =======
Thanks in advance, David -- 389 users mailing list 389-users@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/389-users
-- Carsten Grzemba
-- 389 users mailing list 389-users@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/389-users
389-users@lists.fedoraproject.org
-
Carsten Grzemba -
David Nguyen