Bug 783306

Summary: kerberos auth fails after relocating service to passive node
Product: Red Hat Enterprise Linux 6 Reporter: Nate Straz <nstraz>
Component: opensshAssignee: Jakub Jelen <jjelen>
Status: CLOSED INSUFFICIENT_DATA QA Contact: BaseOS QE Security Team <qe-baseos-security>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 6.2CC: mvadkert, nstraz
Target Milestone: rc   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2017-09-27 12:57:35 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Nate Straz 2012-01-19 21:54:27 UTC
Description of problem:

I have a cluster of two nodes (dash-02 and dash-03) with a virtual IP address (rg-206).  Each node has a keytab with it's principal and the principal for the virtual IP.  I set "GSSAPIStrictAcceptorCheck no" in /etc/ssh/sshd_config on both nodes so I can ssh to the virtual IP's name and still authenticate with Kerberos.

This all works correctly when I set up and start authenticating.

[root@dash-01 tmp]# kinit testmonkey
Password for testmonkey:
[root@dash-01 tmp]# klist
Ticket cache: FILE:/tmp/krb5cc_0
Default principal: testmonkey

Valid starting     Expires            Service principal
01/19/12 15:47:04  01/20/12 15:47:04  krbtgt/EXAMPLE.COM
        renew until 01/19/12 15:47:04
[root@dash-01 tmp]# ssh -l testmonkey dash-02 uname -n
dash-02
[root@dash-01 tmp]# ssh -l testmonkey dash-03 uname -n
dash-03
[root@dash-01 tmp]# ssh -l testmonkey rg-206 uname -n
dash-03

When I relocate the service to dash-02 authentication fails to rg-206.

[root@dash-01 tmp]# ssh -l testmonkey rg-206 uname -n
testmonkey@rg-206's password:
dash-02


Version-Release number of selected component (if applicable):
openssh-5.3p1-70.el6.x86_64


How reproducible:
Easily

Steps to Reproduce:
0. Start with three nodes and a virtual IP
1. Create a Kerberos realm on node 1
2. Create a two node cluster on nodes 2 and 3 w/ virtual IP
2. Set "GSSAPIStrictAcceptorCheck no" in /etc/ssh/sshd_config on cluster nodes
3. Copy ssh host keys from node 2 to node 3
4. Create keytab for node 2 including principals for node 2 and virtual IP
5. Create keytab for node 3 including principals for node 3 and virtual IP
6. Create user principal and user on all nodes
7. kinit user
8. ssh -l user virtual IP
9. relocate virtual IP to other node
10. ssh -l user virtual IP


Actual results:
ssh asks for a password on the second node after IP is relocated to it


Expected results:
ssh should not ask for a password regardless of where the virtual IP is located.

Additional info:

Comment 2 RHEL Program Management 2012-05-03 05:18:12 UTC
Since RHEL 6.3 External Beta has begun, and this bug remains
unresolved, it has been rejected as it is not proposed as
exception or blocker.

Red Hat invites you to ask your support representative to
propose this request, if appropriate and relevant, in the
next release of Red Hat Enterprise Linux.

Comment 3 Jakub Jelen 2017-09-13 11:25:55 UTC
Is it still the case for RHEL7/Fedora? This is most probably not going to be fixed in RHEL6, since it was not a big deal during last 5 years, but if it is something that you are interested in or you have a customer for, please, provide a debug logs, that might give us some more information about the issue.

Otherwise I will close this bug.

Comment 4 Jakub Jelen 2017-09-27 12:57:35 UTC
Closing since there was no answer for two weeks. If it is still an issue, please open a new bug for RHEL7/Fedora.