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

[dlandon]

6 minutes ago, Stormblessed21 said:

Yes, as that's where the data that I am attempting to migrate is stored.

It looks like you need to go to the Windows computer and share the complete 'D' drive.  I didn't have mine shared and that's probably why it didn't work for me.  I don't want to share my complete drive, so you'll have to give it a shot.

[MrTyton]

Hello,

 

I seem to be having an issue with remote SMB shares randomly timing out when trying to read data off of them.

 

Specifically, the attachment shows what happens when I try to play a video through plex, where the video is located on the remote share - I start getting `Error... took longer than 5 seconds`, and the display up top starts showing the size as 0 bytes (the screenshot doesn't show the //COCOON_Plex one as being down, but it did go down at the same time as the error message popup, just a skew as to when I took the screenshots)

 

I know that the remote share is online throughout this - I can still access it through my windows machine, and I have another (non-unraid) server on the network that has been playing plex videos through a remote mount on the same device without any problems, it's only on this new unraid server with these remote mounts that there seems to be this disconnect problem.

 

Is there something that I misconfigured, or some way to fix this? Most of the reason that I bought this new machine was to have it host plex, but due to the size of the library I have the actual files on the synology NAS instead of on the local machine.

 

Thanks,

-MrTyton

[sjoerd]

@dlandon @ich777

The problem solved it self after "the windows remedy" (reboot the system) - tried restarting the array first but that didn't cut it. Now I see the drive as Dev3 and the gears-icon is now also orange (instead of white) - And as default the drive is not being shared on mount.

[dlandon]

10 hours ago, MrTyton said:

 

I seem to be having an issue with remote SMB shares randomly timing out when trying to read data off of them.

The timeouts with 'df' are most likely network issues.  The 'df' command is used to get the stats on the remote share - size, used and free space.  The 'df' command is prone to hanging which is why UD times out the command if it is taking too long.

 

Post your diagnostics zip file for further help.

[dlandon]

22 hours ago, Marv said:

So to be able to remove the partition after a preclear with UD I have to zero the MBR with -z.

I've put some time into this issue and have it resolved.  I am using a different utility now to clear a disk and it does remove the preclear signature and partition created by the binhex preclear docker.  The fix will be in the next release.

[Marv]

2 hours ago, dlandon said:

I've put some time into this issue and have it resolved.  I am using a different utility now to clear a disk and it does remove the preclear signature and partition created by the binhex preclear docker.  The fix will be in the next release.

 

nice, thanks. But this means that unraid would preclear the drive again when adding it to the array afterwards right?

[dlandon]

This is not needed for adding a disk to the array.  Once a disk is preclered, it can be added to the array without formatting.  This is in fact the preferred method for adding a new disk.

 

I fixed it because there are users that might want to remove the preclear signature and format a disk for use in UD or on another system.  It needed to be fixed for that.  You shouldn't need to go through several steps to clear the disk.

[chuckthetekkie]

Hopefully that fixes the format issues I was having a while ago where I couldn’t format unless I canceled a preclear after doing a full preclear. Thanks. 

[dlandon]

After you do a full preclear, you need to clear the stats on the preclear status line before the 'Format' button is enabled.  It shows up as 'Stop Preclear', but it removes the stats file once the preclear is finished.

 

You still need to do that step before you can format a disk.

[MrTyton]

12 hours ago, dlandon said:

The timeouts with 'df' are most likely network issues.  The 'df' command is used to get the stats on the remote share - size, used and free space.  The 'df' command is prone to hanging which is why UD times out the command if it is taking too long.

 

Post your diagnostics zip file for further help.

 I ended up just changing it to NFS, which is a bit more of a hassle but no issue. Thanks anyways! Love the plugin.

[b0m541]

This is just an idea I just had when I wanted to unmount a USB disk and the unmount failed.

In such a situation the user can look at the disk log or open a terminal and investigate manually.

Some things the user would probably do is if there are any file handles open on the disk.

I think it would be beneficial if UD would collect such diagnostic information on its own when an unmount fails (using tools such as lsof) and present it to the user.

 

What do you think about this idea?

[dlandon]

13 hours ago, b0m541 said:

This is just an idea I just had when I wanted to unmount a USB disk and the unmount failed.

In such a situation the user can look at the disk log or open a terminal and investigate manually.

Some things the user would probably do is if there are any file handles open on the disk.

I think it would be beneficial if UD would collect such diagnostic information on its own when an unmount fails (using tools such as lsof) and present it to the user.

 

What do you think about this idea?

UD used to show open files on the Gui, but I removed it because UD is already doing a lot of querying for this kind of information and it gets to the point where UD is slowing things down just to collect information.

 

The log entry is very clear about open files.  It states that the device is busy and can't be unmounted.

 

I like the general idea, but haven't exactly come up with a slick way to show a user the problem(s).  There are a lot of reasons for a failed unmount and UD would have to try to sort through that and show a user what's happening (or not happening as it were).

[slimshizn]

On 1/21/2022 at 9:46 PM, Lev said:

 

I had the same experience after updating the UD plugin to the latest version just now. All my UD drives had been moved to history. Thanks for the quick fix you provided, it resolved the situation easily.

Hadn't came back till now, no issues here since, I was just slightly annoyed at the time haha. Glad we all are back to chugging along.

[Marv]

On 2/27/2022 at 6:47 PM, dlandon said:

This is not needed for adding a disk to the array.  Once a disk is preclered, it can be added to the array without formatting.  This is in fact the preferred method for adding a new disk.

 

I fixed it because there are users that might want to remove the preclear signature and format a disk for use in UD or on another system.  It needed to be fixed for that.  You shouldn't need to go through several steps to clear the disk.

 

 

So something is strange for me.

I just precleared another drive with binhex_preclear container.

In UD I know have the top red 'x' to remove all partitions from the disk and mount is greyed out.

So everything as it should be.

I don't want to do anything with UD from here but add the precleared disk to my array.

So I stop the array, add another disk slot and am able to select the specific disk.

But after that unraid does not show the selected disk in the slot but tells me the slot is 'unassigned'.

However, in UD it now shows the greyed out 'array' button...

 

I am able to stop the array again and remove the unassigned disk slot.

After restarting the array everything is like before and UD shows the greyed out 'mount' again aswell as the top red 'x'.

 

I had this same issue with the disk before. That's why I tried removing the partition with UD and posted my issue the other day.

 

I did not clear the precleared partition yet or zeroed the MBR again with binhex_preclear.sh -z ... as it should work just as I tried above right?

 

 

 

 

[dlandon]

2 hours ago, Marv said:

But after that unraid does not show the selected disk in the slot but tells me the slot is 'unassigned'.

However, in UD it now shows the greyed out 'array' button...

This looks like an issue with Unraid assignments.  Post a new issue on the appropriate forum and include your diagnostics.

 

Whether or not the disk is precleared, Unraid should accept the disk and do what it needs to do to include it in the array.

[b0m541]

On 3/1/2022 at 12:44 AM, dlandon said:

UD used to show open files on the Gui, but I removed it because UD is already doing a lot of querying for this kind of information and it gets to the point where UD is slowing things down just to collect information.

 

The log entry is very clear about open files.  It states that the device is busy and can't be unmounted.

 

I like the general idea, but haven't exactly come up with a slick way to show a user the problem(s).  There are a lot of reasons for a failed unmount and UD would have to try to sort through that and show a user what's happening (or not happening as it were).

 

I get your point.

Your aspiration with this is more ambitious than what I actually had in mind:

 

I didn't think that UD is actually analyzing the collected information to pinpoint the actual problem. I was rather interested in the part that is easier to implement and makes it more comfy to people who are able to manually read/interpret and analyze some diagnostic tool output - which would be collected by UD only when the user pressed Unmount and the button was not set to "Mount" by UD. So a one-off situation when UD would collect and present the diag info and definitely not repeatedly or continually.

But of course, as I respect and are grateful that you do all the work, mentioning this idea is all this is from my side. No expectations at all.

Edited March 3, 2022 by b0m541

[nmills3]

I'm trying to mount my nvme drive but the mount button is just greyed out. what am i doing wrong?

[dlandon]

That disk doesn't have a file system.  'FS' is blank.

[nmills3]

I thought that you got a format option if there wasn't a file system. i'm trying to mount the drive so i can use it to test if something is a disk speed issue or not with another application

[dlandon]

Remove the partition and you can then format the disk.  Go to the UD settings and enable destructive mode and then you can remove the partition.

[snailtrails]

I keep getting Unraid OS warnings for SMART attribute 197 on my unassigned drive (SSD). There is a known crucial firmware bug about this so I disabled monitoring of this attribute on the drive's smart setting page yet I still keep getting browser and email notifications about it. Is this an unassigned devices bug or unraid OS bug?

[dlandon]

It comes from the devX designation not being consistent.  Unraid changes the devX at times and because the disk tracking is by devX and not serial number, Unraid re-reports the error thinking it's a change on a new disk.

 

I've put this on the list of bugs in Unraid.

[comet424]

i cant get my ssd to mount.. i getting some ata errors..  i tried my other server and same errors..  is the ssd cooked.   or can i force a mount   etc  any info? i took 2 diagnostics  1  and the 2nd one i tried a different sata

 

i tried  reboot and a shut down and turn back on  but cant access it

as i need to get the files off it

 

3rd one is from my other server..  how can i scan disk like force a mount etc

 

i also getting the other errors now too  thats from version 2022.02.27b... gonna update to your latest

Mar 10 22:14:04 backupserver kernel: BTRFS warning (device <unknown>): duplicate device /dev/sdb1 devid 1 generation 741242 scanned by udevd (3489)
Mar 10 22:14:13 backupserver kernel: ata1.00: failed to IDENTIFY (I/O error, err_mask=0x40)
Mar 10 22:14:29 backupserver kernel: ata1.00: failed to IDENTIFY (I/O error, err_mask=0x40)
Mar 10 22:14:41 backupserver kernel: BTRFS warning (device <unknown>): duplicate device /dev/sdb1 devid 1 generation 741242 scanned by udevd (3609)
Mar 10 22:14:42 backupserver kernel: BTRFS warning (device <unknown>): duplicate device /dev/sdb1 devid 1 generation 741242 scanned by udevd (3609)

tower-diagnostics-20220310-2034.zip tower-diagnostics-20220310-2024.zip

backupserver-diagnostics-20220310-2135.zip

backupserver-diagnostics-20220310-2233.zip

Edited March 11, 2022 by comet424

[comet424]

maybe spinner drives are better for hot swapping on regular basis    since something happened to SSD ugh.. 

as i swap with my sisters server.. it was working i unmount and she slides the ssd out of the tray and adapter tray.. and then i get it.. now i cant mount it..   

[dlandon]

2 hours ago, comet424 said:

 

i also getting the other errors now too  thats from version 2022.02.27b... gonna update to your latest

You're getting a duplicate device ID on /dev/sdb.  Try going to the UD settings and change the UUID on that disk.  It's towards the bottom of the page.