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

[dlandon]

All, Please update UD and see if it solves your issues.

[alturismo]

15 hours ago, dlandon said:

I just remembered there is a trick that should fix this for you without a reboot,  Just add the share as if you were setting it up for the first time and it will update and 'fix' the configuration.

thanks for the hint, actually didnt work out with hostname.

 

i can add the shares "again", either manually or by searching Server, working fine and end with "success", but no change in UI, still red crossed like above and still in the background mounted.

 

message after successfully added again via host name

 

Aug  8 05:30:26 AlsServer unassigned.devices: Remote server 'ALSBATH' cannot be found on the LAN.
Aug  8 05:32:17 AlsServer unassigned.devices: Remote server 'ALSSLEEP' cannot be found on the LAN.

 

while i can ping them

 

root@AlsServer:~# ping alsbath
PING alsbath (192.168.1.45) 56(84) bytes of data.
64 bytes from f4706921-c46a-0527-2f5e-9de96c7e6154.fritz.box (192.168.1.45): icmp_seq=1 ttl=64 time=3.85 ms
64 bytes from f4706921-c46a-0527-2f5e-9de96c7e6154.fritz.box (192.168.1.45): icmp_seq=2 ttl=64 time=5.59 ms
^C
--- alsbath ping statistics ---
2 packets transmitted, 2 received, 0% packet loss, time 1001ms
rtt min/avg/max/mdev = 3.854/4.721/5.588/0.867 ms
root@AlsServer:~# ping alssleep
PING alssleep (192.168.1.25) 56(84) bytes of data.
64 bytes from 6ef73ac1-4a98-69b4-564a-c8d31d4dd96b.fritz.box (192.168.1.25): icmp_seq=1 ttl=64 time=2.83 ms
64 bytes from 6ef73ac1-4a98-69b4-564a-c8d31d4dd96b.fritz.box (192.168.1.25): icmp_seq=2 ttl=64 time=0.882 ms
^C

 

so i thought, well, lets add them again with IP instead name and voila, back again normally in the UI

 

 

thanks for the Tipp with just readding ... worked out in the end

[ziggy99]

3 hours ago, dlandon said:

All, Please update UD and see if it solves your issues.

Works perfectly now after the fix. The path is now better (no more /mnt in it), the scripts are running as expected and historic devices shows scripts nicely.

 

Thanks for quick fix!

[Kilrah]

I just spun up my test unraid usb on an old machine to clear 5 drives that were previously used in a Drobo DAS, hadn't been booted in a while and all 5 drives were available to clear or format, installed all the plugin updates including UD and now 2 of them were shown as "pool". Had do use wipefs -a to clear the "raid member" signature from them.

A bit too cautious maybe? not great if there is no GUI way to clear previously used drives that had a layout that isn't even in use on unraid.

[dlandon]

4 hours ago, Kilrah said:

A bit too cautious maybe? not great if there is no GUI way to clear previously used drives that had a layout that isn't even in use on unraid.

No.  Let me give you an example.  A user has a VM running Linux and passes through a disk for the VM.  The VM formats the disk using LVM without a partition.  Because Unraid disks all have partitions (including UD), this disk layout isn't understood by UD.  UD expects that all disks in Unraid will have a partition.  You'll notice that the device settings icon (three gears) doesn't show unless a disk has a partition, so marking a disk as "Passed Through" is not possible.  The "Pool" status was added so UD would not perform any operations on the disk or let the User perform any operations.  A user just had this situation.  That's why this was implemented.

 

Until I can come up with a way to allow the clearing of a disk in this state that guards against a user accidentally losing data, this is the way it will stay.

[dlandon]

8 hours ago, alturismo said:

i can add the shares "again", either manually or by searching Server, working fine and end with "success", but no change in UI, still red crossed like above and still in the background mounted.

 

message after successfully added again via host name

 

Aug  8 05:30:26 AlsServer unassigned.devices: Remote server 'ALSBATH' cannot be found on the LAN.
Aug  8 05:32:17 AlsServer unassigned.devices: Remote server 'ALSSLEEP' cannot be found on the LAN.

 

while i can ping them

Even though your server can be ping'd, UD could not resolve those two server names to IP addresses.  Did you ping by server name or IP address?

 

As you found out, you can use either the server name or the IP address in UD.  The situation you were in was probably you doing an edit to the config file to get UD to "think" it had resolved the server name.  That used to work, but I've been working through a lot of issues with UD that caused problems with character sets that make php fail.  The Chinese character set has been particularly challenging.  That's why the situation you had failed after a UD update.

[alturismo]

13 minutes ago, dlandon said:

Did you ping by server name or IP address?

to check i used the name of course

 

 

and all is good, just happened on the update when i posted and wondered ... readding by ip resolved it and everything is fine here, thanks for taking a look and the hint with readding

[dlandon]

3 minutes ago, alturismo said:

to check i used the name of course

The reason I asked was to confirm the name lookup.  UD seems to be having difficulty resolving names to IP addresses at times.

[Amr0d]

I am not able to mount my remote smb shares. The remote unraid is online and I can ping the IP but as soon as I click the mount button nothing happens. What could cause this issue?

tower-diagnostics-20230808-2103.zip

[dlandon]

16 minutes ago, Amr0d said:

What could cause this issue?

You have a routing issue:

Aug  3 01:49:58 Tower kernel: CIFS: Attempting to mount \\10.253.0.6\amr0d
Aug  3 01:49:58 Tower kernel: CIFS: VFS: Error connecting to socket. Aborting operation.
Aug  3 01:49:58 Tower kernel: CIFS: VFS: cifs_mount failed w/return code = -111
Aug  3 01:49:58 Tower unassigned.devices: SMB 1.0 mount failed: 'mount error(111): could not connect to 10.253.0.6Unable to find suitable address. '.
Aug  3 01:49:58 Tower unassigned.devices: Share '//10.253.0.6/amr0d' failed to mount.

 

According to the testparm output, your listening interfaces are:

interfaces = 192.168.94.69 10.253.0.1 127.0.0.1

 

Not sure what the problem is because I don't know the configuration of your network. 

[Amr0d]

2 hours ago, dlandon said:

Not sure what the problem is because I don't know the configuration of your network. 

 

192.168.x.x is my local network and the 10.253.0.x is my Wireguard network. The only thing I did was a reboot and from there it didn't work anymore. No changes in my network so far.

[dlandon]

1 minute ago, Amr0d said:

 

192.168.x.x is my local network and the 10.253.0.x is my Wireguard network. The only thing I did was a reboot and from there it didn't work anymore. No changes in my network so far.

Try adding one of the shares back.  You can add the same remote share and it will update the UD settings.

 

I am using a Wireguard connecton to a remote server and it works for me.

[SShadow]

After downloading 2023.08.07 the mount button for a SMB remote share does not change on the GUI.  I have re-created the share and rebooted my server.  Further troubleshooting, the share actually mounts looking at the log and I can get to the remote folder from file manager.
 

Prior to 2023.08.07 and 2023.08.06a everything worked normally.  I could mount directly to a sub-folder in the share for example:  //192.168.0.10/Backup/Unraid

 

Now //192.168.0.10/Backup/Unraid the mount button does not change and there is no size information.

 

If I re-create the share using //192.168.0.10/Backup the mount button works normally and the size information shows.

 

Is this expected behavior with the changes to plugin or a bug?  Let me know if I did not give enough detail.  Thank you

Edited August 9, 2023 by SShadow

[dlandon]

8 hours ago, SShadow said:

Is this expected behavior with the changes to plugin or a bug?

I honestly didn't know that would work.  Some of the changes recently caused this issue for you.  I've made a change that will fix this.  It will be in the next release.

[mgutt]

After updating unassigned devices to 2023.08.07 in Unraid 6.11.5 the whole Unraid GUI got broken:

 

I was forced to delete the plg files from the flash drive and reboot the server. Now after doing a fresh install of the plugin everything seems to be normal. Not sure what the problem caused and sadly I disabled the syslog mirror option 🙈

 

Some other errors, which were visible after the update and disappeared after reboot:

 

Dynamix File Manager

 

Community Applications

 

 

 

 

 

[Wintermuteau]

On 8/10/2023 at 3:07 AM, mgutt said:

After updating unassigned devices to 2023.08.07 in Unraid 6.11.5 the whole Unraid GUI got broken:

 

 

I was forced to delete the plg files from the flash drive and reboot the server. Now after doing a fresh install of the plugin everything seems to be normal.

Have also been having unraid gui issues for the past few days, it will not load for me, and I also tried what mgutt did and it also went back to normal, for ~1 hour, then the gui went on the fritz, not loading/forever loading for me, again.

[dlandon]

1 hour ago, Wintermuteau said:

Have also been having unraid gui issues for the past few days, it will not load for me, and I also tried what mgutt did and it also went back to normal, for ~1 hour, then the gui went on the fritz, not loading/forever loading for me, again.

Several days ago there was an UD preclear release that had some files missing,  That release caused some gui issues.  It was posted for about five minutes.  You may have caught that release.  I'm not sure if that is causing this for you, but remove the UD preclear plugin and see if yur issues clear up.

[Wintermuteau]

On 8/11/2023 at 9:24 PM, dlandon said:

Several days ago there was an UD preclear release that had some files missing,  That release caused some gui issues.  It was posted for about five minutes.  You may have caught that release.  I'm not sure if that is causing this for you, but remove the UD preclear plugin and see if yur issues clear up.

Apologies sir, it was not your plugin, I forgot I turned on IP4 + 6 on the same day, and that seems to be the problem, turn it back to IP4 only, no more problems with the GUI. So all good.

[jimeez]

I updated most of my hardware today (mobo, CPU, memory).  Everything is working fine, however; I can no longer mount my Remote SMB share that resides on a Windows Server 2016 machine.  I can successfully set up the share with username, password, domain, etc. It shows up on the MAIN tab and indicates that it is "online.  But when I click MOUNT, it tries and then just goes back to saying MOUNT again.  I did try replacing the share using the static IP address rather than the name of the server.  No luck. 

 

Many thanks in advance.

[Kilrah]

Did you try removing the $ from the mount point?

[jimeez]

16 minutes ago, Kilrah said:

Did you try removing the $ from the mount point?

Man, do I feel like a dumb a$$ now.  Thank you very much. 

[Kilrah]

Seem to have found an edge case when unmounting with UD.

When clicking the button UD sends a `sync -f mountpoint` which flushes the filesystem, then unmounts and returns "successfully unmounted". It doesn't however sync the device itself after the unmount and wait for / ensure the writes involved in the unmounting itself have been committed before reporting a successful unmount.

 

I've run into the situation of a badly "clogged" SMR drive where for UD it was unmounted well and good but the drive still needed minutes before actually being safe to unplug. Maybe an additional device-level sync would be a good addition.

Edited August 13, 2023 by Kilrah

[dlandon]

19 minutes ago, Kilrah said:

I've run into the situation of a badly "clogged" SMR drive where for UD it was unmounted well and good but the drive still needed minutes before actually being safe to unplug. Maybe an additional device-level sync would be a good addition.

Syncing the disk device itself is not necessary.  Linux handles all buffer flushing when unmounting.  The added 'sync -f ...' is to allow the syncing to occur before attempting an unmount because it can potentially take a long time and the unmount command would time out.

 

When the 'umount' command completes, Linux has completed all it's writes to the disk.  A device level sync will do nothing because the disk writes are already completed.

 

What you are probably running into is the disk itself buffering data and taking a while to finish writing.  My suggestion is to add a delay in your device script to give the appropriate amount of time to be sure the device has indeed finished writing all data.

 

For example:

 'REMOVE' )
    # do your stuff here

	# Let the disk finish it's disk writes before reporting disk is unmounted
	# Wait for five minutes to be sure the disk has completed its write operations
	sleep 300

    /usr/local/emhttp/webGui/scripts/notify -e "Unraid Server Notice" -s "Unassigned Devices" -d "Device unmounted" -i "normal"
  ;;

 

And wait for the notificaton before removing the disk.

[Kilrah]

It's not a device script but a manual unmount, but

 

33 minutes ago, dlandon said:

When the 'umount' command completes, Linux has completed all it's writes to the disk.  A device level sync will do nothing because the disk writes are already completed.

 

What you are probably running into is the disk itself buffering data and taking a while to finish writing.  My suggestion is to add a delay in your device script to give the appropriate amount of time to be sure the device has indeed finished writing all data.

I am pretty certain it was not the case because 

- UD button was back to "mount"

- Syslog has "Successfully unmounted", i.e. the notification would have been received

 

yet:

- Main page still showed writes to the drive

- IOwait was still high

- Unplugging the device (see timestamps, a whole 6 minutes later) caused xfs to go haywire, IOwait dropped confirming it was coming from that drive

- Next mount needed several minutes to replay the log

 

All of these suggest activity on the OS level after unmount, the unmount having "written" to the disk and returned but not synced and those still being slowly flushed from the OS cache...

 

Aug 11 12:42:52 Unraid unassigned.devices: Unmounting partition '/dev/sdc1' at mountpoint '/mnt/disks/WCJ4D6L2'...
Aug 11 12:42:52 Unraid unassigned.devices: Synching file system on '/mnt/disks/WCJ4D6L2'.
Aug 11 12:49:12 Unraid unassigned.devices: Warning: shell_exec(/bin/sync -f '/mnt/disks/WCJ4D6L2') took longer than 90s!
Aug 11 12:49:12 Unraid unassigned.devices: Unmount cmd: /sbin/umount '/mnt/disks/WCJ4D6L2' 2>&1
Aug 11 12:49:12 Unraid unassigned.devices: Successfully unmounted '/dev/sdc1'
Aug 11 13:05:12 Unraid kernel: usb 2-2: USB disconnect, device number 10
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274673912 op 0x1:(WRITE) flags 0x1000 phys_seg 8 prio class 2
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274689336 op 0x1:(WRITE) flags 0x1000 phys_seg 8 prio class 2
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274696760 op 0x1:(WRITE) flags 0x1000 phys_seg 8 prio class 2
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274704248 op 0x1:(WRITE) flags 0x1000 phys_seg 8 prio class 2
Aug 11 13:05:12 Unraid kernel: XFS (sdc1): metadata I/O error in "xfs_buf_ioend+0x111/0x384 [xfs]" at daddr 0x879500f8 len 32 error 19
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274714680 op 0x1:(WRITE) flags 0x1000 phys_seg 8 prio class 2
Aug 11 13:05:12 Unraid kernel: XFS (sdc1): metadata I/O error in "xfs_buf_ioend+0x111/0x384 [xfs]" at daddr 0x87951df8 len 32 error 19
Aug 11 13:05:12 Unraid kernel: XFS (sdc1): metadata I/O error in "xfs_buf_ioend+0x111/0x384 [xfs]" at daddr 0x87953b38 len 32 error 19
Aug 11 13:05:12 Unraid kernel: XFS (sdc1): metadata I/O error in "xfs_buf_ioend+0x111/0x384 [xfs]" at daddr 0x879563f8 len 32 error 19
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274689336 op 0x1:(WRITE) flags 0x1000 phys_seg 4 prio class 2
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274689368 op 0x1:(WRITE) flags 0x1000 phys_seg 4 prio class 2
Aug 11 13:05:12 Unraid kernel: device offline error, dev sdc, sector 2274696760 op 0x1:(WRITE) flags 0x1000 phys_seg 4 prio class 2
Aug 11 13:05:12 Unraid kernel: XFS (sdc1): Metadata I/O Error (0x1) detected at xfs_buf_ioend+0x251/0x384 [xfs] (fs/xfs/xfs_buf.c:1260).  Shutting down filesystem.
Aug 11 13:05:12 Unraid kernel: XFS (sdc1): Please unmount the filesystem and rectify the problem(s)

 

Edited August 13, 2023 by Kilrah

[dlandon]

1 hour ago, Kilrah said:

Aug 11 12:49:12 Unraid unassigned.devices: Warning: shell_exec(/bin/sync -f '/mnt/disks/WCJ4D6L2') took longer than 90s!

This is the problem.  The sync command timed out.  I ill have to add more time to the sync timeout.

 

EDIT: I'll set it up so the sync will not time out in this situation, but the sync timing out didn't necessarily cause the problem of writing to the disk after it is unmounted.  If the sync is allowed to complete, it might help, but Linux finished the unmount and supposedly flushed out the disk cache before ending the unmount.