Description of problem: I have a collection of RHL 9 boxes in a lab environment. I am exporting /home and /var/spool/mail over nfs, using the following export file: &/root& # cat /etc/exports /var/ftp/pub *(ro,insecure,async,all_squash) /var/ftp/pub 130.160.20.0/255.255.255.0(rw,insecure,async,no_root_squash) /home *.unix.eng.ua.edu(rw,insecure,async,no_root_squash) /home 130.160.20.0/255.255.255.0(rw,no_root_squash,insecure,async) /var/spool/mail *.unix.eng.ua.edu(rw,insecure,async,no_root_squash) /var/spool/mail 130.160.20.0/255.255.255.0(rw,insecure,async,no_root_squash) /home/student/crutcher monket.dyndns.org(rw,insecure,async) After applying the latest set of updates, to client and servers, I am getting the following error, intermittently, on mount attempts: # On the client side: [root@ho-l06 root]# mount /home mount: home.unix.eng.ua.edu:/home failed, reason given by server: Permission denied # From /var/log/messages on the server Jun 11 04:23:06 mule rpc.mountd: authenticated mount request from ho-l06.unix.eng.ua.edu:877 for /home (/home) Jun 11 04:23:06 mule rpc.mountd: getfh failed: Operation not permitted These machines are all in the unix.eng.ua.edu domain, and all use the following hosts.allow file: &/var/log& # cat /etc/hosts.allow # # hosts.allow This file describes the names of the hosts which are # allowed to use the local INET services, as decided # by the '/usr/sbin/tcpd' server. # statd: .unix.eng.ua.edu mountd: .unix.eng.ua.edu The clients use the following fstab file: LABEL=/ / ext3 defaults 1 1 LABEL=/boot /boot ext3 defaults 1 2 none /dev/pts devpts gid=5,mode=620 0 0 none /proc proc defaults 0 0 none /dev/shm tmpfs defaults 0 0 /dev/hda3 swap swap defaults 0 0 /dev/cdrom /mnt/cdrom udf,iso9660 noauto,owner,kudzu,ro 0 0 /dev/fd0 /mnt/floppy auto noauto,owner,kudzu 0 0 home.unix.eng.ua.edu:/home /home nfs async,rsize=8192,wsize=8192 0 2 home.unix.eng.ua.edu:/var/spool/mail /var/spool/mail nfs async,rsize=8192,wsize=8192 0 2 There is NO /etc/sysconfig/nfs file on these systems. nfs and nfslock are configured to run, and status reports that they are. These machines are current as of this posting with RHL9 and all updates, applied through redhat network on the account 'rhnunival'. There are still some old RHL8.0 profiles in that account, as I am in the middle of a migration that is taking entirely too long. This worked before applying the updates. I am very concerned about this, I have tried many things, but have identified very few common threads. First, after a reboot, the server seems to permit mounts to home for a while. Then it stops. Once it stops, it never starts again. However, I have seen no similar problems with /var/spool/mail mounts. I have searched bugzilla for relevent bugs, and found nothing that pertains to this problem. I am willing and able to do sersious debuging in my environment, if I am given direction on how to do it. I do not possess NFS debugging knowledge, or I would have given a better report here. I will be forced to roll back updates on the server tomorrow, in an effort to find a functional state for the system. At this point even a work-arround would do, my computer network does not work at all right now, locking out about 150 college students from their linux accounts.
This is some kind of interaction with nfs and symlinks. I had some external locations (/var/ftp and /var/spool/mail) linked to locations in, and between these being exported, and the links, NFS was getting confused. Since a client box would mount /home FIRST, then /var/spool/mail, it worked the FIRST time. Setting no_subtree_checks in the exports file didn't help, though. Still having Stale NFS filehandle bugs, but that is a different issue. Closing this, I guess.