Red Hat Bugzilla – Bug 530591
Can't start nfs.statd in F12Beta LiveCD
Last modified: 2009-11-20 08:42:49 EST
I booted a F12Beta LiveCD, and needed to mount an NFS share (on an F11 machine). I discovered that nfs-utils was not installed, so I did this with yum. I then tried to start nfs.statd, and it failed. I was then only able to mount the NFS share using the -o nolock option.
I'll be happy to provide any additional info needed.
Is SELinux enabled? If so, try doing as root
'setenforce && service nfslock start'
It is. Assuming you really meant "setenforce 0 && service nfslock start", that fails.
/var/log/messages gives "unable to register (statd, 1, udp)."
Ah so rpcbind is not running... what does rpcinfo -p output?
rpcinfo: can't contact portmapper: RPC: Remote system error - no such file or directory
Ok.. that is the problem... rpcbind is not is not started which means
rpc.statd will not start (by design) which means only '-o nolocks' mount
So the question is why is rpcbind not being started?? Unfortunately
I'm not sure to whom I ask that question...
I'd normally say the rpcbind maintainer, but that looks like it's you. :) And I don't see anything earth-shattering between the version in F-11 and the rawhide one.
I installed the Fedora12 beta livecd, now fully updated and wanted to set up an NFS share with my laptop running Fedora 7 (laptop only has 256mb ram, it's old, so it's probably staying with F7). When I tried to mount the exported folder on the laptop, I got a message 'permission denied'. I opened the ports I set up in /etc/sysconfig/nfs on the firewall -- so it shouldn't actually be a permission issue.
I thought I'd then try the reverse, to see if I could mount a folder to the F12 pc, and here's the output:
/usr/sbin/start-statd: line 8: /sbin/rpc.statd: Permission denied
/usr/sbin/start-statd: line 8: /sbin/rpc.statd: Success
mount.nfs: rpc.statd is not running but is required for remote locking.
mount.nfs: Either use '-o nolock' to keep locks local, or start statd.
mount.nfs: an incorrect mount option was specified
I ran the command from message 2 "setenforce 0 && service nfslock start", then did: 'cat /var/log/messages |grep statd' and got:
Nov 7 18:27:44 localhost rpc.statd: Version 1.2.0 Starting
Nov 7 22:31:56 localhost rpc.statd: Caught signal 15, un-registering and exiting.
Nov 8 09:04:08 maple rpc.statd: Version 1.2.0 Starting
Then ran 'nfs service restart' and now I can mount the folder on the F12 machine without the -o nolock option, but I still get a 'permission denied' trying to mount a folder on the laptop. Let me know if you need any more info.
Just to add to the above comment 7; on the /var/log/messages, the one with today's date (Nov 9, 09:04:08) was from when I booted the pc, because it's now 2 hours later, pacific time when I got it to mount without the -o nolock. In case it matters.
I figured it out, I think. I opened up the NFS config to try allowing connections higher than port 1024, in the off chance that made a difference. While I was looking at the 'host' info, I saw there was an 'http:/' in front of the IP number, I didn't put that there.
I looked at /etc/exports, and sure enough it had the host IP prepended by http:/. I edited the /etc/hosts file and removed the http:/ ; did a 'service nfs restart' then I could mount the share on my laptop just fine.
I deleted the http:/ in >system > administration >NFS (a package I'd installed), and when I exited and pulled it up again; I found it now has a 2nd entry with the http:/ added back in again. Where that's coming from I don't know.
This may or may not be related to the rpc.statd issue -- if it's a separate issue I can put this someplace else.
So the export entry in /etc/exports had a 'http:/' on it?
(In reply to comment #10)
> So the export entry in /etc/exports had a 'http:/' on it?
Yes, AdamW suggested (on fedora forums) I file a bug report, here it is:
This bug appears to have been reported against 'rawhide' during the Fedora 12 development cycle.
Changing version to '12'.
More information and reason for this action is here:
*** This bug has been marked as a duplicate of bug 533771 ***