[HTCondor-users] Submit fails - RHEL 7 in NIS environment and FS authentication

["Feldt, Andrew N." <afeldt@xxxxxx>]

Folks,

Does anyone run RHEL 7 with NIS and attempt to use FS authentication with Condor?  I have been trying to set up a test system and I have Condor starting properly, but when I try to submit a job from this system, I get:

ERROR: Failed to connect to local queue manager
AUTHENTICATE:1003:Failed to authenticate with any method
AUTHENTICATE:1004:Failed to authenticate using FS

I turned on debugging (and restricted authentication to only FS) via the following config:

SEC_DEFAULT_AUTHENTICATION_METHODS = FS
SCHEDD_DEBUG = $(SCHEDD_DEBUG) D_FULLDEBUG D_SECURITY
TOOL_DEBUG = $(TOOL_DEBUG) D_FULLDEBUG D_SECURITY

and, I then find the following in the SchedLog file:

06/30/15 11:00:52 (pid:3046) FS: client template is /tmp/FS_XXXXXXXXX
06/30/15 11:00:52 (pid:3046) FS: client filename is /tmp/FS_XXXGS7c9Q
06/30/15 11:00:52 (pid:3046) AUTHENTICATE_FS: used dir /tmp/FS_XXXGS7c9Q, status: 0
06/30/15 11:00:52 (pid:3046) AUTHENTICATE: method 4 (FS) failed.
â
06/30/15 11:00:52 (pid:3046) DC_AUTHENTICATE: reason for authentication failure: AUTHENTICATE:1003:Failed to authenticate with any method|AUTHENTICATE:1004:Failed to authenticate using FS|FS:1004:Unable to lstat(/tmp/FS_XXXGS7c9Q)

This is for a valid, logged in user in my NIS password map.  There are other entries later for the local Condor user (which is in /etc/passwd) which show that it authenticates just fine with FS authentication:

06/30/15 11:57:05 (pid:3046) FS: client template is /tmp/FS_XXXXXXXXX
06/30/15 11:57:05 (pid:3046) FS: client filename is /tmp/FS_XXXmNulOS
06/30/15 11:57:05 (pid:3046) AUTHENTICATE_FS: used dir /tmp/FS_XXXmNulOS, status: 1
06/30/15 11:57:05 (pid:3046) AUTHENTICATE: auth_status == 4 (FS)
06/30/15 11:57:05 (pid:3046) Authentication was a Success.
06/30/15 11:57:05 (pid:3046) ZKM: setting default map to condor@xxxxxxxxxx
06/30/15 11:57:05 (pid:3046) ZKM: post-map: current user is 'condor'
06/30/15 11:57:05 (pid:3046) ZKM: post-map: current domain is 'nhn.ou.edu'
06/30/15 11:57:05 (pid:3046) ZKM: post-map: current FQU is 'condor@xxxxxxxxxx'
06/30/15 11:57:05 (pid:3046) DC_AUTHENTICATE: authentication of 129.15.31.37 complete.

So, local users (which my ordinary users are not) work fine, but NIS users (which all my ordinary users are) fail.

I have tried running nscd (which I have to run on my RHEL 6 systems to let Condor function), but this does not help.

I would be happy to hear any suggestions!

Andy