[evanpurkhiser]

Hey guys,

So just yesterday my XBMC mysteriously stopped being able to connect to an NFS share that's been working fine for the past few months..

I get the error "RPC Packet not accepted by the server" in my xbmc.log

Here's the relavant section from my xbmc.log and my /etc/exports

I've also tried running XBMC as root to see if that would work, but it results in the same problem.

I've also tested to see how generic the error is, if I give XBMC an invalid NFS path it will just say "failed", giving it a valid path produces the error above.

I'm really stumped by this one so any help would be appreciated. Googling the error only seems to turn up post related to the OSX version of the NFS server.

[Kib]

I will not even look at that log because i know it is a log snippet instead of a full debug log.
*READ THIS 1st BEFORE POSTING* FAQs, HOW-TO post about a problem in a useful manner

[evanpurkhiser]

Sorry about that =)

Here is the full xbmc.log with debugging enabled: https://gist.github.com/EvanPurkhiser/5607489

That's from reproducing the error immediately after a clean start.

Also, here's are some additional details about the system.

XBMC is running on a Arch Linux x64 (Linux 3.9.2-1-ARCH) with all up to date packages. XBMC is v12.2 (installed with arches package manager 'pacman').

The NFS server (desktop) is also on Arch Linux x64 (Linux 3.9.2-1-ARCH) with all up to date packages.

[evanpurkhiser]

I've just installed XBMC on the same machine hosting the NFS server and am able to reproduce the very same problem. (Just to be clear, I updated my /etc/exports to allow world access to the share and ran exportfs -af).

Also, the nfs server appears to be authenticating the connection just fine, in my system log (journalctl) I see the line

Code:

May 19 07:53:45 desktop rpc.mountd[10386]: authenticated mount request from 192.168.1.101:50050 for /home/evan/multimedia (/home/evan/multimedia)

[evanpurkhiser]

Downgrading from Linux 3.9.2 => 3.8.8 solves this.

I'm looking into what changed in the kernel that causes this now.

I'm not going to mark this thread as solved just yet, downgrading the kernel is just putting a bandaid on the wound.

Edit: I've opened a thread on the Arch Linux forum to hopefully get some insight into why this breaks in 3.9.2

[evanpurkhiser]

Just installed XBMC nightly (from herecura testing repo http://repo.herecura.be/), the problem still persists in that build.

I've also upgraded the NFS server to Linux 3.9.3 since it came out last week, this did not solve the problem either.

I've also tried adding a NFS share on the HTPC (which is also running 3.9.3) and connecting to that on XBMC. Once again, I can connect fine from the CLI, but not through

Is there anyone out there who is running a NFS server on the Linux kernel > 3.8.8 that is NOT experiencing this problem?

I'm still not entirely sure if this is a regression in the linux kernel, or if a configuration option changed that I've missed.

Not sure if I should open a ticket in Trac for this issue or not. I really can't tell if it's related to XMBC, libnfs, or the Linux kernel itself.

Still looking for help going forward on this one. I'm holding back the kernel on my desktop, but I'm starting to have to hold back some other packages that require linux >= 3.9. Another way I could bandaid this is to add the NFS share to my fstab and just add the path as a source in XBMC.

[Memphiz]

You Could add an issue on github.com/sahlberg/libnfs and ask if it is a Known issue

[evanpurkhiser]

Sounds good. I will! Thanks!

Done!

[evanpurkhiser]

This has been fixed in 1.6 of libnfs =)

[Memphiz]

Ok will bump version in XBMC to 1.6 then.

[Memphiz]

follow PR2793

[evanpurkhiser]

Exactly one year later and I'm having this exact (or very very similar) same problem again

Looks like someone already got to it first this time. Doesn't appear to be a kernel issue this time, also haven't managed to get it back to a working state by rolling things back. But I'm suspecting it is the same issues as what's mentioned in that Arch forum topic.

If anyone else has seen this I wouldn't mind a little help

Edit: It isn't the same error though, this time a more generic "nfs_opendir_async failed". Actually feel a little silly bumping this thread now. Sorry about that!

[Memphiz]

You don't need to feel silly about bumping an old thread. Its up to you how you feel about not posting a debug log (wiki) though

[evanpurkhiser]

Here is the xbmc.log and the logs for the rpc.mountd service on my NFS server (Relevant line is 594 in the log)

Downgrading linux versions isn't appearing to fix it this time around.