RHEL6: /sbin/mount.nfs hangs with a backtrace in lookup_root when attempting to mount a share from RHEL5.11 NFS server
Issue
We have an automounter config mounting some remote directories via NFS. Occasionally, the mount process hangs for one of the directories. The process that actually hangs is /sbin/mount.nfs. The automounter configuration for that directory is no different from the configuration for other directories, and other directories don't hang.
- Specifically the /foo/mnt is hanging during mount but not all times. When it succeeds it gets mounted as NFSv3 but the mount seems to be hanging at NFSv4 operation.
Environment
-
Red Hat Enterprise Linux 6 (NFS client)
- seen on RHEL6.9
- kernel-2.6.32-696.13.2.el6.x86_64
- autofs-5.0.5-132.el6.x86_64 Sat May 13 11:22:37 2017
- nfs-utils-1.2.3-75.el6.x86_64 Sat May 13 11:21:34 2017
- nfs-utils-lib-1.1.5-13.el6.x86_64 Sat May 13 11:21:33 2017
-
Centrify is being used
- The /etc/auto.master file has an entry like this, where /etc/auto.customscript is a perl script from Centrify
/foo /etc/auto.customscript
When /foo/mnt is accessed, the script is run and the autofs maps are in Active Directory which gets queried for the map. The actual map that pops out of all that looks like this and points at an export on a RHEL5 NFS server
mnt rhel5.example.org:/export
- Red Hat Enterprise Linux 5 (NFS server)
- seen on RHEL5.11
- kernel-2.6.18-419.el5
- nfs-utils-1.0.9-71.el5_11.x86_64
- nfs-utils-lib-1.0.8-7.9.el5.x86_64
- exports
/export *.dom1.example.org(rw,sync,no_root_squash,fsid=5) host1.example.org(rw,sync,no_root_squash,fsid=5) host2.example.org(rw,sync,no_root_squash,fsid=5) host3.example.org(rw,sync,no_root_squash,fsid=5)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.