Red Hat Bugzilla – Bug 786982
Mount hangs with sec=krb5 option.
Last modified: 2012-03-13 14:39:38 EDT
Description of problem:
When using the -o sec=krb5 option, the mount command hangs.
Version-Release number of selected component (if applicable):
Fedora release 14 (Laughlin)
Easy, happens everytime.
Steps to Reproduce:
1. kill rpc.gssd daemon (ps -aef | grep rpc.gssd, kill -2 the PID).
2. Try a mount with the given option:
mount -o sec=krb5 nfsserver:/ /mnt/nfsserver/
The command hangs for ever.
It should error out saying the rpc gss is not running.
See the /var/log/messages file which tells us to re-start rpc.gssd.
This sounds more like a mount or nfs change request rather than one for the libraries, so I'm moving this to nfs-utils, though it might end up in kernel space.
Note, though, that F14 has been past its end-of-life for several weeks, so it's worth knowing if this is something that still happens on a release where this could be fixed: F15 or F16.
I have observed with fed14, fed13, RHEL 6.1.
(In reply to comment #2)
> I have observed with fed14, fed13, RHEL 6.1.
There was a check but it got removed in upstream because it was to distro specific... The same goes for the idmapper check.
The checks consisted of checking for the existence of the /var/lock/subsys/<daemon> file. With the introduction of systemd, that file is no longer created when these daemons startup...
While it was nice to have those checks I just to see a simple way to extract the status out of systemd without doing a fork/exec(), which I think is a bit overkill.