Unassigned Devices - Managing Disk Drives and Remote Shares Outside of The Unraid Array

[dlandon]

Just now, btrcp2000 said:

Same here, no impact from clearing cache

Read previous post.

[dlandon]

4 hours ago, PlayerOne said:

Nothing in the syslog at all.

Is it a password protected share?

[btrcp2000]

A little progress. I dumped FreeNFS and installed Hanewin as an NFS Server on my Win10 VM, although I'm not sure that matters now. I apparently had a firewall issue on the WinVM, now fixed, so within UD, typing in the VM's IP and clicking Load Shares now finds the folder.  It is also visible to my UbuntuVM and a standalone Win10 PC with NFS Client running.  Neither UD nor Ubuntu will mount it though, which I think has something to do with syntax because I have little idea what I'm actually doing. I have a share setup on Unraid called r5000 (export=yes, security=public for both NFS and SMB). Here's what UD log shows:

 

 

Same error from Ubuntu VM (for testing only, I do not plan to 

 

[John_M]

You have to specify the complete path to the exported share, not just the share name - at least that's the case with Unix-like operating systems. I don't know how a Windows NFS server converts between Widows and Unix paths though. Your Ubuntu mount command is looking for a share called /r5000 - i.e. in the "root" of the Windows file system, wherever that is. Your UD mount command is looking for it at //c/r5000. Neither is correct.

[John_M]

Try 192.168.1.138:/mnt/user/r5000 or 192.168.1.138:/c/mnt/user/r5000 in the first instance. If neither helps the documentation for your NFS server might shed some light.

[btrcp2000]

 

The UD mount command is what UNRAID found when I clicked load shares, i didn't enter it like that.  No matter what I put in the mount point it reverts to this:  (this was because i wasn't hitting enter before clicking mount).

 

now i see this in the log:

 

but /mnt/disks/r5000 definitely exists:

 

and UNRAID can see the win10 NFS share from its CLI

 

what am I not seeing?

 

 

Edited October 28, 2017 by btrcp2000

[dlandon]

1 hour ago, btrcp2000 said:

 

The UD mount command is what UNRAID found when I clicked load shares, i didn't enter it like that.  No matter what I put in the mount point it reverts to this:  (this was because i wasn't hitting enter before clicking mount).

 

now i see this in the log:

 

but /mnt/disks/r5000 definitely exists:

 

and UNRAID can see the win10 NFS share from its CLI

 

what am I not seeing?

 

 

You seem to be making this way too complicated.  Enter the IP address in the 'IP/Host' field - 192.168.1.138, click on the 'Load Shares' button and select the share you want.  UD will take care of the rest.

 

If that produces errors or the NFS share doesn't mount, show a screen shot of what you are entering into the dialog for IP address and share, and the UD log so I can help troubleshoot.  If you are entering all kinds of syntax trying to get the NFS share to mount, it is difficult to understand and hard to see what UD might be doing wrong.

[btrcp2000]

I removed the existing share to start over, entered the IP ending 138. "Load Shares" found the folder as shown:

 

 

Clicking "Add" results in this without any edits to the mount point:

 

 

Clicking Mount results in a "Mounting" but it never completes. Log (also attached) looks like this:

 

unraid-unassigned.devices-20171027-2106.zip

[SG872]

Using your plugin I found I can only mount one of my 4 drives:

Two of them show ZFS member (I experimented with FreeNAS at one point but rebuilt the array and everything was hunky dory)

What I am trying to do is recover some data from the disks I have removed (rebuilt onto 2x8TB) as they couldn't be read from backups and TBH not sure which disk the data was on.  I think sde was my parity disk, if that is the case would that explain me being unable to mount it?  Any idea what I could possibly do with these zfs member disks?

[dlandon]

7 hours ago, btrcp2000 said:

I removed the existing share to start over, entered the IP ending 138. "Load Shares" found the folder as shown:

 

 

Clicking "Add" results in this without any edits to the mount point:

 

 

Clicking Mount results in a "Mounting" but it never completes. Log (also attached) looks like this:

 

unraid-unassigned.devices-20171027-2106.zip

The 'No such file or directory' message is because the server is off-line and the rc5000 file cannot be found.  Here is a part of the UD log:

Oct 26 17:25:36 Error: SMB share '192.168.1.138:/mnt/user/r5000://mnt/user/r5000' appears to be off-line.
Oct 26 17:25:36 Error: Remote SMB/NFS server '192.168.1.138:/mnt/user/r5000' is offline and share '192.168.1.138:/mnt/user/r5000://mnt/user/r5000' cannot be mounted.

UD pings the server at 192.168.1.138 to see if it responds to check if the server is on-line before mounting the share.  You need to see why the 192.168.1.138 IP address does not respond to a ping.

 

There are several issues I see here with UD.  I made some changes to try to help with shares that go off-line by logging those events.  The log entry is showing the wrong information though.  The other issue is UD should not even try to mount a share that is not on-line.

 

I'm going to work on cleaning up this situation.

[btrcp2000]

I ssh'd into unraid as root and had it ping 138, it responded as expected.  The only folder that exists now on the Nfs server is c:/r5000.  The entry from 17:25 is from when I actually had a share called /mnt/user/r5000 on the win VM to test. I had removed it from the Nfs server but not yet from UD, so it was still looking for it when the log was taken. 

 

Appreciate the effort, and apologies if I'm a step or two behind. I am a *nix late bloomer

[dlandon]

7 hours ago, SG872 said:

Using your plugin I found I can only mount one of my 4 drives:

Two of them show ZFS member (I experimented with FreeNAS at one point but rebuilt the array and everything was hunky dory)

What I am trying to do is recover some data from the disks I have removed (rebuilt onto 2x8TB) as they couldn't be read from backups and TBH not sure which disk the data was on.  I think sde was my parity disk, if that is the case would that explain me being unable to mount it?  Any idea what I could possibly do with these zfs member disks?

I am not a zfs expert, but it appears those two zfs member disks were mirrors in a zfs setup.  UD can't mount zfs disks.  You will need to take them to another system that can work with zfs disks and remove the data from there.  I know there is a zfs plugin for unRAID, but I have no idea how it works.

 

I found this on mounting zfs member disks: https://www.svennd.be/mount-unknown-filesystem-type-zfs_member/ that might help.

 

Show the UD log for the xfs disk that won't mount after trying to do a mount.

[dlandon]

6 minutes ago, btrcp2000 said:

I ssh'd into unraid as root and had it ping 138, it responded as expected.  The only folder that exists now on the Nfs server is c:/r5000.  The entry from 17:25 is from when I actually had a share called /mnt/user/r5000 on the win VM to test. I had removed it from the Nfs server but not yet from UD, so it was still looking for it when the log was taken. 

 

Appreciate the effort, and apologies if I'm a step or two behind. I am a *nix late bloomer

I am getting even more confused.  What is it that you are trying to do?  If I recall your posts, you are trying to mount a NFS share in a VM with UD?  Why are you doing that?

[btrcp2000]

I have a TV tuner for sagetv that can only run on windows. The sagetv docker controls it but it instructs it to record programs to a Linux directory, which windows can't see. I have found one other person on the planet who has gotten this to work, but they haven't responded, so I'm trying to recreate(see here)

 

Goal is to dump my standalone windows sagetv server and use unraid for everything, but I'm stuck with the r5000 tuner as I am under contract. If this can't be done I will have to abandon sagetv docker and leave it on the windows VM. Not the end of the world but now that I'm getting more comfortable with *nix I like windows less and less

[dlandon]

15 minutes ago, btrcp2000 said:

I have a TV tuner for sagetv that can only run on windows. The sagetv docker controls it but it instructs it to record programs to a Linux directory, which windows can't see. I have found one other person on the planet who has gotten this to work, but they haven't responded, so I'm trying to recreate(see here)

 

Goal is to dump my standalone windows sagetv server and use unraid for everything, but I'm stuck with the r5000 tuner as I am under contract. If this can't be done I will have to abandon sagetv docker and leave it on the windows VM. Not the end of the world but now that I'm getting more comfortable with *nix I like windows less and less

It seems to me you could set up an unRAID share that is both SMB (Windows can see it) and NFS (sagetv can see it).  Then pass the share through to the sagetv docker.

[btrcp2000]

Agreed, but in my mind that's what I'm trying to do. Sagetv docker passes through a Linux path to the Windows software when it tells it what and where to record. Windows complains it can't find the path. So I need Linux to tell windows to use a path both can see. If I understood the person on the link I sent correctly, they created an nfs share on Windows and mounted it to a Linux server (although not unraid), then adjusted the sagetv properties a certain way. I'm very comfortable with that last part but cannot figure out the shares. 

 

I am very open to any other ideas and would have given up by now if I hadn't discovered a possible solution in the link. Also at the point where I would be willing to pay someone to stand this up. 

[dlandon]

7 minutes ago, btrcp2000 said:

Agreed, but in my mind that's what I'm trying to do. Sagetv docker passes through a Linux path to the Windows software when it tells it what and where to record. Windows complains it can't find the path. So I need Linux to tell windows to use a path both can see. If I understood the person on the link I sent correctly, they created an nfs share on Windows and mounted it to a Linux server (although not unraid), then adjusted the sagetv properties a certain way. I'm very comfortable with that last part but cannot figure out the shares. 

 

I am very open to any other ideas and would have given up by now if I hadn't discovered a possible solution in the link. Also at the point where I would be willing to pay someone to stand this up. 

You don't need UD to mount the NFS share.  It can all be done through unRAID shares.  I don't totally get how the shares all need to work, but unRAID share mapping can do it for you.

[Squid]

I have a problem with UD running on 6.4.0-rc10b

 

It will not mount any of my SMB mounts that are hosted on my secondary unRaid server (6.3.5)

Oct 28 08:40:29 Mount SMB/NFS command: mount -t cifs -o rw,nounix,iocharset=utf8,_netdev,file_mode=0777,dir_mode=0777,username=andrew,password=******* '//SERVER_B/Movies' '/mnt/disks/SERVER_B_Movies'
Oct 28 08:40:29 Mount of '//SERVER_B/Movies' failed. Error message: mount error(95): Operation not supported
Refer to the mount.cifs(8) manual page (e.g. man mount.cifs)

I reverted back to 6.4.0-rc9f for testing, and there are zero problems mounting those shares.

 

[dlandon]

23 minutes ago, Squid said:

I have a problem with UD running on 6.4.0-rc10b

 

It will not mount any of my SMB mounts that are hosted on my secondary unRaid server (6.3.5)

Oct 28 08:40:29 Mount SMB/NFS command: mount -t cifs -o rw,nounix,iocharset=utf8,_netdev,file_mode=0777,dir_mode=0777,username=andrew,password=******* '//SERVER_B/Movies' '/mnt/disks/SERVER_B_Movies'
Oct 28 08:40:29 Mount of '//SERVER_B/Movies' failed. Error message: mount error(95): Operation not supported
Refer to the mount.cifs(8) manual page (e.g. man mount.cifs)

I reverted back to 6.4.0-rc9f for testing, and there are zero problems mounting those shares.

 

Works fine for me.  Samba was updated in rc10b.  I think this is related to the samba version on the remote (security) and what version the remote wants to use.

 

I have an idea about how to get around this with UD.  It is coming up more and more.  Can I PM you and have you do a few tests for me?

[Squid]

1 minute ago, dlandon said:

Can I PM you and have you do a few tests for me?

Knock yourself out

[dlandon]

3 minutes ago, Squid said:

Knock yourself out

I'll do some testing here first and then have you do some tests.  I don't have a remote share that fails like you are seeing to test.

[btrcp2000]

1 hour ago, dlandon said:

You don't need UD to mount the NFS share.  It can all be done through unRAID shares.  I don't totally get how the shares all need to work, but unRAID share mapping can do it for you.

 

If this is what you're saying, I think I've tried the "vice versa" approach before, having the Unraid share be the NFS server, and using windows as the NFS client, so I just tried again.  This is what the windows VM (192.168.1.138) shows. Unraid is 192.168.1.100 

 

If I edit the sagetv docker config file to show R:\r5000 as the recording drive, upon startup of the docker it gets changed to /opt/sagetv/server/R\:\r5000 which the windows software of course can't find

 

Just throwing stuff at the wall to see what sticks:

 

Recording drive as \\192.168.1.100\mnt\user\r5000 prevents the docker from even starting

 

/mnt/user/r5000 lets the docker start but it complains that the drive is out of space upon recording (it's not)

 

This link is where all this began almost three weeks ago, listing all the stuff i've tried. 

 

 

 

 

 

 

 

[dlandon]

33 minutes ago, btrcp2000 said:

 

If this is what you're saying, I think I've tried the "vice versa" approach before, having the Unraid share be the NFS server, and using windows as the NFS client, so I just tried again.  This is what the windows VM (192.168.1.138) shows. Unraid is 192.168.1.100 

 

If I edit the sagetv docker config file to show R:\r5000 as the recording drive, upon startup of the docker it gets changed to /opt/sagetv/server/R\:\r5000 which the windows software of course can't find

 

Just throwing stuff at the wall to see what sticks:

 

Recording drive as \\192.168.1.100\mnt\user\r5000 prevents the docker from even starting

 

/mnt/user/r5000 lets the docker start but it complains that the drive is out of space upon recording (it's not)

 

This link is where all this began almost three weeks ago, listing all the stuff i've tried. 

 

 

 

 

 

 

 

You can pass the /mnt/user/r5000 to the docker directly with read/write privileges.  Use "Add another path, port, or variable" in the docker template and map the share as a path specifying the unRAID share as the Host Path and the container path.

[btrcp2000]

That was one of the first things I did:

 

Back to having Unraid mount an outside NFS share, for whatever it's worth, here are the settings on the other end which again is a Win10 VM hosted on the Unraid server, using hanewin as the NFS server. Firewall is completely off for testing purposes, and I have tried checking/unchecking "Allow mount of remote devices":

 

 

Edited October 28, 2017 by btrcp2000

[dlandon]

3 hours ago, Squid said:

I have a problem with UD running on 6.4.0-rc10b

 

It will not mount any of my SMB mounts that are hosted on my secondary unRaid server (6.3.5)

Oct 28 08:40:29 Mount SMB/NFS command: mount -t cifs -o rw,nounix,iocharset=utf8,_netdev,file_mode=0777,dir_mode=0777,username=andrew,password=******* '//SERVER_B/Movies' '/mnt/disks/SERVER_B_Movies'
Oct 28 08:40:29 Mount of '//SERVER_B/Movies' failed. Error message: mount error(95): Operation not supported
Refer to the mount.cifs(8) manual page (e.g. man mount.cifs)

I reverted back to 6.4.0-rc9f for testing, and there are zero problems mounting those shares.

 

Here is the situation.  It is all related to samba.  Samba used to default to the SMB1 protocol which all devices would understand.  I suspect for security reasons, the default seems to be SMB3.  I am implementing a scheme where the first cifs mount attempt will be SMB3, then if that fails SMB2, and then finally SMB1.  This will mount the SMB share at the highest protocol the remote server will allow.