steve1977

Here we go, thanks for your help! tower-diagnostics-20251228-0720.zip

Let me clarify. Everything works as long as I use the Tailscale IP. I can access my unraid server remotely via the Tailscale IP. What no longer works is to access it remotely via the local network IP. When on my local network, this of course still works as well. So assume must be somehow related to either the plugin update breaking this functionality or one of the plugin settings (subnet routing? dns?).

Have been a loyal tailscale plugin user for a while. It really works so well. Thank you for the development! I am not sure whether it is related to a recent update, me toying around with magic-dns, or any setting change (which I don't believe is the case). Any thoughts appreciated! When I an using the tailscale IP, all works well and I can access the Unraid UI over tailscale. Unfortunately, I can no longer access via my local network IP (192.168.x.x) since recent. I am still advertising the IP, so this hasn't changed. Not sure what else I have done that prevents this from working now?

Thanks for detailed write-up. I am but confused though. You say that you suggest NFS shares, but you use SMB shares yourself? Why is this? Let me try to recap your setup Unraid Server #1 (home) Appdata with arr dockers, Plex docker [unshared] Appdata with makemkv docker [shared] ZFS pool for media Unraid Server #2 (remote) Appdata with makemkv docker [using mapped appdata from unraid server #1, only read access]

Got it, this makes sense to me. So could even be a good deal. Same as you, I also still have a 120GB SSD, which must be one of the very first generations and must have cost a fortune. It obviously is totally outdated technology by now, but I keep using it for time to time to try out new things. For example, I just thought about using it to convert one of my VMs to a barebone install. The diskspeed docker doesn't work to test the PM981 as it doesn't work for ZFS pools. But it seems for this exchange that it may well be an authentic - though very old - disk.

It was sold as "new", but who knows... As you said, I wasn't overly suspicious as it is old, old technology, and OEM. I was looking to buy 2.5" SATA SSD disks and realized how expensive they still are although equally oudated technology. Was surprised the SATA disks are as expensive as modern NVME disks. That triggered my thinking that the PM981 disks were too cheap to be authentic. Well, maybe just got lucky. I just installed diskspeed docker and giving this a shot. If this passes at solid speed, I'd assume they are just old and potentially refurbished of some form. But at least not fake. And given I am using it as a ZFS mirror disk, the worst is that it fails and then I just replace it.

This was not exactly what I was looking for, but also seems very interesting. Are you saying you’re sharing the docker config directories across two unraid servers? One only has read access to avoid data corruption? How does it work with “watched flags”?

I bought an absurdly cheap nvme from eBay. It’s labeled as pm981 1TB, which is meant to be some Samsung oem disk. I’m currently using it as zfs mirror disk. It seems to do the trick and haven’t had any issue so far. I’m wondering though whether it’s even real and not some fake product. Is there any way to check out whether the disk is faulty / speed / reliable / etc? Often things to good to be true are just faulty…

Thanks for the detailed write-up. I’m interested in using unraid as Time Machine backup. A few thoughts: I always thought that sparse bundles don’t play well with xfs? I remember I copied one to an array and this turned into a total mess. I thought APF formats was the way to go? Is this no longer an issue or does ZFS play nicer with TM sparsebundles? Did you ever try to mirror the disk? I’m thinking to make this disk the backup for my most sensitive data (several TM backups, personal documents). Besides mirroring the disk, could I also sync it with it iCloud? Could I also have onedrive, iCloud and other local files on this zfs disk? Thanks for any thoughts you may have?

Any update on this and what are the limitations of mandaron’s docker?

Did you try it out and dos this still work in 2025/26?

Title says it all. Any ideas from the diagnostics? tower-diagnostics-20251224-0630.zip

It seems I may have celebrated too early. Even after changing from user to disk share, the Plex music library scan is not completing. It did run much longer though, so this may be by chance? Attaching the dianostic log, but the issue may be a Plex docker issue rather than an Unraid issue? Thoughts? tower-diagnostics-20251220-2048.zip

Fair. Let me do that. Let me also keep monitoring whether it indeed works reliably with disk shares. Will keep monitoring and revert with logs.

I am facing an issue with Plex, but I don't think it is a Plex issue per se. So, better posting here. When adding my music library via user share to Plex, the scan isn't reliable. It stops every few minutes and I cannot scan all of it as I need to restart every few minutes. It all works very well if I add the library via a disk share (instead of user share). So, this is an issue about the reliability to access data via a user share instead of disk share. Any thoughts whether there is some Unraid tuning that can make user share faster / more reliable / more accessible?

Unfortunately not. It seems that wifi in its current state is still too hackery, so I gave up on it. Hope that it will eventually be fully supported in the future.

Thanks. Got it working. Eventually the plugin upgrade became available. No idea why it didn't do, but after a few hours and a dozen "refreshs", it showed up. Now all good.

Unraid says that I need to upgrade the Connect plugin, but Connect says it is already the latest version: Current Unraid: 7.2.2 Connect plugin: 2025.11.21.0223 plugin: installing: unRAIDServer.plg Executing hook script: pre_plugin_checks plugin: downloading: unRAIDServer.plg ... done Executing hook script: pre_plugin_checks *** *** Please update the Unraid Connect plugin before upgrading Unraid! *** plugin: run failed: '/bin/bash' returned 1 Executing hook script: post_plugin_checks

I have two Unraid servers (2 licenses). Both are running on different networks. I am wondering whether I can have docker containers access each other? This would require me to mount the array from unraid server #1 as disks on unraid server #2. Would this be possible? Maybe through UD?

I am curious whether it is possible to connect two Unraid servers in different locations over Tailscale. And then access their respective arrays inside their dockers (docker in unraid server a accesses docker of unraid server b). This would require me to mount the unraid a array in unraid b?

Got it. This explains. The work-around works.

total 190 drw-rw-rw- 11 nobody users 11 Nov 23 12:06 ./ drwxrwxrwx 7 nobody users 8 Nov 24 08:04 ../ drw-rw-rw- 8 nobody users 8 Nov 14 23:12 .removed/ drw-rw-rw- 3 nobody users 3 Nov 17 22:19 5\ removed drw-rw-rw- 2 myusername users 16 Nov 23 11:55 removed\ -\ Album\ -\ 2025\ -\ removed/ drw-rw-rw- 3 myusername users 3 Nov 23 08:04 removed/ drw-rw-rw- 3 myusername users 3 Nov 23 08:04 removed/ drw-rw-rw- 3 myusername users 3 Nov 23 08:04 removed/ drw-rw-rw- 3 myusername users 3 Nov 23 08:07 removed/ drw-rw-rw- 3 myusername users 3 Nov 23 08:05 removed/ drw-rw-rw- 3 myusername users 3 Nov 23 08:05 removed/

Thanks. Have done it multiple times. Please see screenshots below:

Thanks. I had applied, but just redid everything and yet again "applied". Below new diagnostic logs. What will have changed is the samba sharing (which I enabled since), but I still cannot accesss the files in Windows and the permissions are still inconsistent. The files are still there though as I can access them via the Unraid GUI. tower-diagnostics-20251124-0803.zip