Bug 607103 - NFSv4 & rpcgssd fails when run as service; works properly when run in foreground
Summary: NFSv4 & rpcgssd fails when run as service; works properly when run in foreground
Keywords:
Status: CLOSED NOTABUG
Alias: None
Product: Fedora
Classification: Fedora
Component: nfs-utils
Version: 13
Hardware: i686
OS: Linux
low
high
Target Milestone: ---
Assignee: Steve Dickson
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2010-06-23 08:45 UTC by Anthony Messina
Modified: 2010-10-15 14:58 UTC (History)
2 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
Environment:
Last Closed: 2010-10-15 14:58:08 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
Failure as service (5.38 KB, application/octet-stream)
2010-06-23 08:45 UTC, Anthony Messina
no flags Details
Success in the foreground (984 bytes, application/octet-stream)
2010-06-23 08:46 UTC, Anthony Messina
no flags Details

Description Anthony Messina 2010-06-23 08:45:18 UTC
Created attachment 426201 [details]
Failure as service

Description of problem:
Note that this only occurs on my i686 machines, not my x86_64 machines.  

Using:
kernel-PAE-2.6.33.5-124.fc13.i686
nfs-utils-1.2.2-2.fc13.i686
nfs-utils-lib-1.1.5-1.fc13.i686

The rpcgssd service starts normally and is visibly running (ps ax|grep rpc shows (rpc.gssd -vvv").  However, on the i686 machines, the user is not able to access nfs4 mounts with sec=krb5p as the the server's nfs/<host> service ticket is not propagated to the user's credential cache.

This occurs with "regular" nfs4 mounts from the command line, from fstab at boot or from autofs.

I've attached a success and failure log.

Comment 1 Anthony Messina 2010-06-23 08:46:00 UTC
Created attachment 426202 [details]
Success in the foreground

Comment 2 Anthony Messina 2010-06-27 08:27:27 UTC
I found this link:

http://www.listshow.net/201005/freeipa-users/28184-freeipa-users-nfs4-after-client-upgrade-to-fedora-13.html

which suggested doing "semanage boolean -1 allow_gssd_read_tmp".

That seems to work.  Sorry about the report --my x86_64 machines were running in permissive mode at the time while the i686 ones were enforcing.

However, It still doesn't make sense that rpc.gssd worked in the foreground in enforcing mode with that boolean set to false??

Comment 3 Steve Dickson 2010-10-15 14:58:08 UTC
> However, It still doesn't make sense that rpc.gssd worked in the foreground in
> enforcing mode with that boolean set to false??
Well alot of the SElinux-izms don't make sense, IMHO... ;-) 


I'm going got ahead a close this, please feel free to reopen


Note You need to log in before you can comment on or make changes to this bug.