Try :

 

mount -t nfs -o lookupcache=none,noac 192.168.0.10:/mnt/user/TVRecordings /storage/recordings

 

In the spirit of keeping JoeL happy

 

PeterB has already suggested you might have better staiblity trying an earlier RC.

 

For me, the stale nfs file handle problem is a lot less severe in rc10.

 

I'd be interested to know whether your experience is the same, if you revert to rc10.

 

That would be the path I'd be looking to take if NFS is important. As a side issue if you're using openelec is there a reason you're not ducking the issue altogether and using samba / smb / cifs instead of nfs?

 

 

 

 

Thanks yet again

Your proposed command gives:

 

mount: invalid number 'none'

 

But, as before, at least does not take me into the BusyBox thing and so is another step in the right direction

 

And - thanks - I do intend to try a roll back to rc.10 at the weekend but since I run a few plug ins etc I am always a tiny bit anxious about rolling backwards (Like to keep as up-to-date as possible)

 

And - thanks again - re SMB suggestion. When I first go in to OpenElec/XBMC I was told that NFS shares alway work better and have never quite been bright enough to argue with this and so have stuck to NFS shares through thick and thin! I think that some of the lesser powered boxes used for XBMC work better with NFS perhaps?

 

If moving to SMB from NFS would make the problem go away then I could probably do that just for my TVRecordings share perhaps ... would it be the same command as the original but just substitute SMB for NFS?

 

Dave

 

And - thanks again - re SMB suggestion. When I first go in to OpenElec/XBMC I was told that NFS shares alway work better and have never quite been bright enough to argue with this and so have stuck to NFS shares through thick and thin! I think that some of the lesser powered boxes used for XBMC work better with NFS perhaps?

 

If moving to SMB from NFS would make the problem go away then I could probably do that just for my TVRecordings share perhaps ... would it be the same command as the original but just substitute SMB for NFS?

 

I run openelec and use smb. I'd love to use NFS but for non openelec purposes. I don't see any reason to avoid smb with an openelec client. You are recording TV though so maximum throughput / minimal latency *theoretically* available with nfs may be more important.

 

Try it. It would certainly avoid the nfs stale file handle problem as you're not using nfs!

 

I don't use the tv options so I don't know how you would switch to smb. I would suggest trying to mount it using the openelec internal setup. You don't really need to 'mount' smb as such - openelec should be able to just go directly to the smb share as necessary based on the openelec internal config.

 

I struggled to find a nice guide on this, not least for tv but this might set you in the right direction :

 

http://xbmc.us/njbetzen/2012/08/01/beginner-tip-how-to-add-videos-to-xbmc-frodopre-frodo-from-a-network-share/

 

Or if you want to use smb in the same way you're trying to manage nfs this looks like it should help :

 

http://wiki.openelec.tv/index.php?title=Mounting_network_shares

Thanks yet again for being soooooo helpful

Now have a few things to try and will report back when I have a working solution.

 

1) Will either change to recording TV to a mounted SMB unRaid share in my OpenElec/XBMC box (the rest of the OpenElec box works perfectly all of the time .... but, of course, it is simply streaming .mkv and .mp3 files from my unRaid server... recording TV is different - I am streaming FROM OpenElec TO unRaid in this case... my HD tuner stick is in my OpenElec Box)

 

or

 

2) Roll back to rc.10 (may do this for the sake of experimentation/help others who may be having or may have the same issue in the future) 

 

Will keep you updated - might help somebody else in the future. Now that TVHeadEnd is working so well in OpenElec I am confident that what I want to achieve will become part of most peoples XMBC installs eventually.

 

Cheers!

Er..... where do I find the earlier unRaid releases now that the site has been "recreated"?

need 5 rc 10

 

Er..... where do I find the earlier unRaid releases now that the site has been "recreated"?

need 5 rc 10

Here

Thanks for this link. Have rolled back to rc10 and so far no Stale NFS File Handles .. been running for about 2 hours now and so under rc12 I would probably have had an error by now.

Will check again tomorrow - have left a few tv programmes recording to this mounted share - and update again

Now 12 hours since I rolled back to rc.10 and no Stale NFS File Handles ... and I've probably recorded around 5 or 6 TV programmes to the mounted cached share... and unRaid has moved from cache to share successfully too. So - at the moment rc.12 seems to cause the issue. What now? Don't like to be behind the latest version - any chance that rc.13 can fix this? Meanwhile any log files etc needed to help in solving this issue?

Will update again after another 12 - 24 hours

Cheers

... any chance that rc.13 can fix this?

 

I've been trying to get Tom to comment on this - stale nfs file handles had been a significant problem for quite some time.  Then, in rc4 (iirc) Tom implemented a change in shfs which effected a considerable improvement.  It appears as though this shfs modification was lost again in rc11.  Current behaviour feels very similar to pre-rc4 behaviour.

I'm slightly confused, it's obvious 12a introduced an issue  with stale NFS so why not just undo the change and incorporate into RC13?

 

Sent from my GT-I9300 using Tapatalk 2

 

 

I'm slightly confused, it's obvious 12a introduced an issue  with stale NFS so why not just undo the change and incorporate into RC13?

 

Actually, it was rc11 which (re-)introduced the issue.  If the problem is not caused by the shfs modification getting lost, it may be that it was introduced with a later Linux kernel which, itself, was updated to address other issues.

hmm tricky, so rolling back to earlier kernel could introduce issues that were fixed in the later kernel, but if tom doesnt roll back the kernel then he has issues with stale nfs handles instead, hello mr rock meet mr hard place :-).

So - for anyone who is interested/following this thread .... over 24 hours have now passed ... and I've recorded over 20 TV Programmes (I know that there is not enough good stuff on to justify 20+recordings in 24 hours.... I am just soak testing the process) and it is working like a dream. No Stale NFS File Handles.

Using rc10

The command that I am using in my XBMC/OpenElec autostart is (in case anyone else is attempting the same setup):

mount -t nfs [my XBMC IP Number]:/mnt/user/TVRecordings /storage/recordings -o nolock;\

Thanks for all the advice - fingers crossed the problem with go away again in rc13?

Dave K

To add a data point:

 

I have recently set up an unRAID server with the intention of mounting two user shares (Movies & Television) to an ubuntu 12.04 system via NFS.

 

I am unable to perform a scan in Plex scan of the two mounted shares on the Ubuntu system without encountering "missing files"/stale NFS file handles.

 

I have rolled back to rc10 and have tried mounting with:

- no options to mount

- defaults,soft,nolock,nfsvers=3

 

Both failed.

 

I am currently testing the following options and will update with success/failure:

- defaults,soft,nolock,nfsvers=3,lookupcache=none,noac

 

I just wanted to post this in case anyone saw davekeel's successful posts and got overly optimistic that a rollback to rc10 would fix all encounters of the NFS issue.

 

I'm really hoping this gets sorted in rc13 and in the very near future. Seriously bummed out at not being able to use my new unRAID install due to this issue.

As an update to my above post:

 

The combination of rc10 and the NFS options of defaults,soft,nolock,nfsvers=3,lookupcache=none,noac specified in my fstab seem to be stable after 24 hours of adding/moving/renaming/scanning content.

The mounting system is Ubuntu 12.04.

 

The addition of 'lookupcache=none,noac' to 'defaults,soft,nolock,nfsvers=3' stabilized the mount immediately. Prior to that option I would start throwing stale handles from Plex scanning the mount.

 

Would still love to see the issues in rc11/rc12/rc12a resolved, but happy to be able to use the system at the moment.

Anyone know whether Stale NFS file handles problems have gone away in rc15a? Still stuck on rc10 but getting itchy to move upwards!?

There have been no reports of this issue under version 15a. And unless someone reports an issue this will be the final version. I suggest you try it and report if you have the issue before it's too late...

I just had one of these.. on rc15, but I havent tried 15a yet. Will do after current pre-clearing is done

Ran a load of TV recordings on my OpenElec XBMC device this weekend (I record these to a mounted unRaid NFS share on Cache) and not a single Stale NFS File Handle error for me (on 15a).

Might not be everyone's experience but, for me, it works fine - if I move back to, say, anything after rc10 then I get stale NFS's errors straight away!!!

Cheers