I recently setup an unraid box. Originally I was using samba exports to copy all my data over to the machine once I got it all setup, but that was painfully slow. I switched to NFS, which sort of worked, but I've been plagued with issues ever since.

 

On the linux machine I can see the exported shares and I can mount them. I can ls the directories, but it's very slow the first time (I assume that's drives spinning up since each other time is fairly quick). I could rsync files from one directory to the nfs export on the unraid box, but it would at random times just stop and I would have to cancel and restart the rsync. If I attempt to cp a file from the nfs export to /tmp it just hangs and doesn't copy any of the data. I have a strace of that I will attach.

 

On the mac it's pretty much the same. From the commandline I can mount the shares, ls the shares, but can not copy anything from the shares. From the GUI I can not even access the shares to view the contents, it just hangs and I have to relaunch finder. When I mounted the shares on the mac, I did with the -p and without (as other posts suggested) but neither made a difference. I also could not get the samba shares to allow me to access any data when I had them exported from samba. I could see the data via finder, but when I attempted to play a video or open a file the application (vlc, textedit) would simply crash.

 

Here are my exports

"/mnt/user/Apps" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=100 *(rw,insecure)
"/mnt/user/Books" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=101 *(rw,insecure)
"/mnt/user/Games" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=102 *(rw,insecure)
"/mnt/user/ISO" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=103 *(rw,insecure)
"/mnt/user/Misc" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=104 *(rw,insecure)
"/mnt/user/Movies" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=105 *(rw,insecure)
"/mnt/user/Music" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=106 *(rw,insecure)
"/mnt/user/PV" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=107 *(rw,insecure)
"/mnt/user/Personal" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=108 *(rw,insecure)
"/mnt/user/TV" -async,no_subtree_check,anongid=0,anonuid=0,all_squash,fsid=109 *(rw,insecure)

 

I have tried with and without the insecure option and neither machine reacts differently.

 

These are the messages from last night and this morning from the linux machine I have the shares mounted on

[root@integra ~]# grep alucard /var/log/messages
Jan  9 17:19:38 integra kernel: [437704.416090] nfs: server alucard not responding, still trying
Jan  9 17:29:21 integra kernel: [438287.328817] nfs: server alucard OK
Jan  9 17:52:47 integra kernel: [439693.415479] NFS: server alucard error: fileid changed
Jan  9 17:53:44 integra kernel: [439750.710992] NFS: server alucard error: fileid changed
Jan  9 17:53:50 integra kernel: [439755.974654] NFS: server alucard error: fileid changed
Jan  9 17:53:50 integra kernel: [439755.975352] NFS: server alucard error: fileid changed
Jan  9 17:53:51 integra kernel: [439757.348345] NFS: server alucard error: fileid changed
Jan 10 09:38:56 integra kernel: [39328.736058] nfs: server alucard not responding, still trying

 

Any help is appreciated. Thanks for reading. On the upside, my data is very securely protected, even I can't manipulate it!

 

I'm not sure if this is relevant but here's some nfs information from unraid and my linux machine:

 

unRAID Logs:

Jan 10 13:36:39 alucard mountd[6312]: authenticated unmount request from 172.16.1.42:634 for /mnt/user/Movies (/mnt/user/Movies)
Jan 10 13:36:59 alucard kernel: svc: 172.16.1.42, port=897: unknown version (4 for prog 100003, nfsd)
Jan 10 13:36:59 alucard mountd[6312]: authenticated mount request from 172.16.1.42:1017 for /mnt/user/Movies (/mnt/user/Movies)

root@alucard:~# rpcinfo -u localhost nfs
program 100003 version 2 ready and waiting
program 100003 version 3 ready and waiting

[root@integra ~]# rpcinfo -u localhost nfs
program 100003 version 2 ready and waiting
program 100003 version 3 ready and waiting

 

Both the unraid machine (alucard) and the workstation (integra) are only using nfs2,3.

 

My setup:

unRAID 4.5.6

MSI 870-G45 AM3 AMD 770 ATX AMD Motherboard

syslog.txt

strace.txt

Version 4.6 changes the nfs mounts are handled. Delete the nfs shares from your clients and the server. Upgrade to 4.6. Reenable the nfs shares and recreate the mounts on your clients.

Thanks.

 

I'll give that a try.

I have upgraded to 4.6.

 

Things started out working exactly as I expect it would, but then took a dump.

 

I was attempting to index the music using gnump3d for streaming and after a while it just stopped indexing (watched a strace and it stops a few moments in). I then tried to simply cp a mp3 from the nfs mounted share to /tmp and it hangs as well. I see this in the messages on the unraid server

 

Jan 10 14:45:31 alucard kernel: rpc-srv/tcp: nfsd: sent only 155780 when sending 524420 bytes - shutting down socket
Jan 10 14:45:34 alucard kernel: rpc-srv/tcp: nfsd: sent only 299140 when sending 524420 bytes - shutting down socket
Jan 10 14:46:14 alucard kernel: rpc-srv/tcp: nfsd: sent only 397444 when sending 524420 bytes - shutting down socket

 

I'm not too sure what that means. When I try to ls that directory now it just hangs as well. Now I'm back to the same problems I had prior to the upgrade. When I bounce both machines, I can copy one file from the nfs mount to local /tmp directory, but the next time I try to copy a file it hangs and all future ls or tab completion operations are very slow to complete. What danger do I encounter if I wipe the flash drive, put 4.6 on there clean and then replace the devices in the same roles they were previously in? Just start with a fresh install? Do I have a high risk of losing data?

Did you completely eliminate all nfs share configs on you client and the server and the recreate from scratch?

 

You could try adding this to your mount command "wrsize=32768"

Ya, I am new to this unraid stuff, but I made sure nothing had any nfs options and it wasn't exporting anything.

Configure OS X nfs mounts in Disk Utility under the file menu. Add wrsize=32768 to the Advance Mount Parameters. I have all of this added to my Advanced Mount Parameters: nolocks, locallocks, intr, soft, wrsize=32768, nosuid

 

You may not need them all.

I've about given up on my unraid server sadly.

 

I have the exact same problem on Fedora, Ubuntu, and even a rescuecd that has nfs available. I can not get this machine to work on anything at all. I'm going to reformat my USB key and put 4.6 on there clean, if that doesn't work I don't know what to do.

I have wiped my flash drive, put 4.6 on fresh. I have made no configuration changes other than adding my drives and enabling user shares and marking 1 share *(rw).

 

This is what I'm seeing

 

[root@integra The.Last.Hero.In.China]# time ls

last hero bak.JPG  The Last Hero In China.avi  The_Last_Hero_In_China-cd.jpg  The_Last_Hero_In_China-front.jpg

 

real 0m11.591s

user 0m0.001s

sys 0m0.007s

 

11 seconds to ls a folder

 

An ls of the top level Movies directory just hangs. After about a minute and a half I can cancel the ls, then I am able to cd quickly into another directory and cat a file. Then it started working rather quickly again when I went through changing directories and cat'ing files. Upon changing directories again, and copying another file to /tmp .. it hangs.

 

I don't know what else to try. I'm completely open to any suggestion.

 

 

Well, have you upgraded the BIOS to the latest very recent version? Try this one first to see.

 

Second - I do not see a problem with the syslog per se but your motherboard is using Atheros chipset - and some of them do not play nice in Unraid. If you have a spare NIC (Intel or Realtek based) you can try to disable the onboard NIC and used the other one to see what will happen.

 

Good luck.

Popped in a crappy ol' netgear nic (back when gigabit nic's were new) and it seems to be doing the trick. Darn that AR8131M chipset.

 

Does anyone know why this is a problem? Like what are the limitations or differences in this chipset vs anything realtek? Is it simply a matter of what's available in slackware at the time?