bmartino1

Calling rsync form apple or rsync form unraid terminal?

I assume a smb connection on the Apple device. as it’s not really a Unraid/XFS bug, it’s mostly about how macOS implements SMB + rsync

Here’s what’s going on:
On macOS, rsync doesn’t talk to the filesystem directly when you’re writing to a remote SMB share — it’s writing through Apple’s SMB client.
Apple’s SMB client is very aggressive with write caching and metadata updates. By default, when rsync finishes copying a file, the mtime (modification time) isn’t guaranteed to be flushed down to disk correctly.
That’s why you see it reset to “now” on the Unraid side instead of being preserved.
When you add --fsync, rsync tells the OS to explicitly flush each file’s data and metadata to disk before moving on. That forces macOS to commit the proper mtime, so you see the expected behavior in Unraid.

Why this is Apple-specific
On Linux rsync over SMB/NFS/SSH, you generally don’t need --fsync.
On macOS, Apple’s SMB stack often ignores utimes()/futimens() metadata updates until the file is fully closed and flushed. Without fsync, the SMB server (Unraid in this case) just gets “best-effort” timestamps.
That’s why it doesn’t matter whether you use rsync 2.6.9 or 3.4.1 — the problem is lower in the stack (SMB + APFS caching behavior).

some potential workarounds...

If you want stable results without always remembering --fsync:
rsync -a --fsync source/ /Volumes/UnraidShare/dest/

Force rsync to go over SSH instead of SMB
Mount your Unraid share via SSH (or just run rsync directly over ssh://).
That way rsync is talking directly to Unraid’s filesystem, bypassing Apple’s SMB client.
rsync -a -e ssh source/ user@unraid:/mnt/user/share/

Use --inplace in some workflows (can help with timestamp consistency if files are being updated instead of replaced).

Check your Unraid SMB settings
Sometimes forcing fruit:metadata=netatalk or disabling AAPL extensions in smb-extra.conf changes how timestamps are handled.
But for your case, it seems squarely an rsync + macOS SMB caching thing.
Rewrite unraids SMB: https://forums.unraid.net/topic/178033-bmartino1-user-scripts/#findComment-1478661

so, just to clarify so I'm understanding you correctly... as it looks like you messed up a special disk without informing zfs of the edit or changes... see warning at the end!

What Happened
You had a RAIDZ2 pool (orion) and attached an SSD as a special read-cache device.
When you physically swapped that SSD and manually removed it from the pool, ZFS left stale metadata device references in the pool config.
Now zpool import shows two pools with the same name:
Healthy pool (ID 6418…) → imports fine, all files accessible.
Broken ghost pool (ID 2517…) → marked UNAVAIL with “corrupted data / insufficient replicas” because the missing special vdev destroyed its internal config.
The ghost pool is confusing Unraid’s UI, which expects only one importable orion.

Some Key Observations
Your data pool is intact — proven by the fact you can zpool import 641826… orion and access /mnt/orion.
The invalid pool is just a stale configuration entry, not another real set of disks. It references corrupted/nonexistent metadata replicas.
The Unraid UI is choking because it sees both references.

so lets review a potential recover plan. as we may need to invoke parts of the new config to erase the old ID and what unraid is tryign to do....

first lets Identify the Good Pool. you should know/have this given that you can mount the zfs array.
zpool import
Note the good pool ID (641826…). That’s the one to keep...
as not all data is in the diag file. the daig help but i don't have enoth info to give full commands.

Next we need to Force Import the Good Pool
If it’s not already imported:
zpool import -f -o cachefile=/etc/zfs/zpool.cache 64182623996842760 orion
*If the 6#### id is the good one...

This writes a fresh zpool.cache containing only the working config.
As we willb e dealign with zfs acl and cofig first...

Then you will need to Clear the Ghost Pool Reference
The “bad” pool is only metadata. To drop it:
Make sure the good pool is imported.
zpool import -D
Should show destroyed pools or corrupted entries
If the bad ID appears, you can remove its ghost reference with
zpool import -D -f <bad_pool_id>
Or, safer: let the system rebuild the cache file (next step).
as the system may do a scrub or resync...

Finally we fix zfs with a cache rebuild with the new disk.
Rebuild ZFS Cache Sometimes Unraid/ZoL just holds onto the stale entry. Refresh the cache:

zpool set cachefile=/etc/zfs/zpool.cache orion

or

zpool export orion
zpool import -o cachefile=/etc/zfs/zpool.cache orion

This ensures only the good pool is seen.
Once the sync is finished...

now we can look at unraid and its UI...
Reboot after the above so Unraid reloads the single valid pool.
If the UI still shows broken references, remove /boot/config/plugins/zfs/zpool.cache (Unraid keeps its own copy) and let it regenerate after next import.
then on the next boot the array should work as intended...
Try the zpool export / import sequence with a rebuilt zpool.cache, then reboot Unraid. That should leave you with only the healthy pool (ID 6418…).
Warnings!!!
Do NOT attempt to import the corrupted pool (2517…). That pool truly has missing devices (special vdev) and importing it risks confusion.
Do NOT run zpool clear or zpool replace on the bad entry. It won’t fix it — it’s not a real pool anymore.
Your data is safe in the good pool. The goal is just cleaning up stale metadata so Unraid stops seeing two.

see guides post for CA and Compose

Docker Guides:
Immich Docker-compose Setup
Immich Unraid-Template in the CA

Please ask questions here.
looks like you went Compose route...
you will need to install the unraid CA Postgres docker and restore your previous postgress database. then update immich to the docker appdata path for immich.
*once you restore the database, the previous immich passord will be needed to connect to the database...

SO, essential install the immich postgress docker I made in the CA.
Run your tools and commands to restore your postgress.
stop the CA docker and add your compose file editing the path to reuse the CA docker data...

-You are responsible for your own data!

contact support

they can assit wil aquireing new lciense for purcahsed products.

you may have to many IP suing the same mac address use macvlan:
Some Docker networking Info when to use what docker network

sorry, your SOL once formatted unless using very expensive tools and software. (the cost is not worth it) the data is gone after a format.

Since this is disk not ssd / nvme cell the data technically still exists on disk in a low level area and technically data can be recovered but not easily and not free... You are better off taking the loss of data in this situation.

What happens when you format?
Formatting a disk essentially resets the file system, which is like a table of contents for the disk, making the data inaccessible through normal means. 
kinda... (imagine that there is a small ssd chip on the HDD disk controler telling the PC how and where to go to find this data... this chip gets erased when you format...)
However, the data itself is usually not immediately erased and remains on the disk until new data overwrites it. 
A quick format only erases the file system, while a full format may overwrite the data with zeros, making recovery more difficult. 

cache is a general pc term meanind to put temperaly before saving...

There are quite a few "cache" things.

there memory cach, there disk cache, there IO Cache the Virtual Cache... would need more context then just the word cache...

Unraid docs would help here:
Storage Management | Unraid Docs
Storage Management | Unraid Docs
To assign devices toShare settings as well...
https://docs.unraid.net/unraid-os/manual/shares/
https://docs.unraid.net/unraid-os/manual/shares/user-shares/
https://docs.unraid.net/unraid-os/manual/shares/disk-shares/

How to explain....

Imagine your Unraid server is a house.
The Array is your big, slow filing cabinet—great for long-term storage, but it takes a while to open drawers.
The Cache is your kitchen counter—fast and easy to reach. You drop things there first before later moving them to the filing cabinet (the Array). Unraid does this automatically using the mover.
A Pool is like an extra room in your house. It's not the kitchen counter (cache), and it's not the filing cabinet (array). It's just another space, and you decide what it's used for.

So, what's the difference between a Cache and a Pool?
Technically, they're both BTRFS/ZFS disk groups—what Unraid now calls pools. The only real difference is whether you assign a pool to act as a cache for the Array (via share settings like "Use cache: Yes/Only/Prefer").

...

So, when is a pool not a cache?
When you're using it for:
Appdata (for Docker/VMs) that should never move to the array
Downloads (Deluge/qBittorrent/SABnzbd etc.) that benefit from SSD speed
Media (e.g. 4K videos for Plex) that you don't want slowed down by the Array
Backups, scratch disks, or ISO libraries that don't need parity

But what about parity for pools?
Pools don’t have Unraid-style single parity.
But BTRFS pools can be mirrored (RAID1), striped (RAID0), or more (RAID5/6 with caution).
Want redundancy? Use RAID1 or RAID10 for your pool.
Want performance? Use RAID0 (no redundancy).

Summary...
Cache = fast temporary space for the array.
Pool = fast permanent space you can use however you want.
Both are “pools,” just used differently. Cache is a role, not a thing.

depends on how its setup but yes this uses the md driver for software raid with mdadm

Unassigned disk plugin?
make a nfs share for a disk  ...

Otherwise...
Under the Main Share of the folder you are using in share make the nfs share private.
and use rule:
192.168.1.x(rw,sec=sys,no_subtree_check) Per your requiem net of removing the no root squash option.... so set the ip of the one machine that is rsyncing file too...

However I will argue that, that seems incorrect on how to access and maintain files over nfs... Nfs doesn't care of the remotes systems file permission and structure unless told otherwise... 
 
Example:
on unraid you will need to chmod and chown the permission for the share before nfs options are set...
Set permission example 
chmod 777 -R * chown nobody:users -R *

But you would want the no root squash for chown of files...
no_root_squash is critical for rsync to be able to chown as root on the remote box.
 
SO, working with Unraid and NFS you would want to configure an NFS share to be:
-Private
-Accessible by root from another Linux machine
-With no_root_squash so rsync can preserve ownership using chown
-Set only through the Unraid web UI

this sounds like a mis-understanding on how the remote machine and how to connect to the share...
as this private rule should work with Rsync...
 
192.168.2.253(rw,sec=sys,no_subtree_check,no_root_squash)

If using /etc/fstab on the client, a line might look like:
unraid:/mnt/user/PBS /mnt/pbs-backup nfs rw,vers=3,hard,intr 0 0
or issues with how your using the rsync command...

...

review:
 
 


would need more data / info to assit.

ip of remote machine....
Command / fstab of remote machine how are you connecting the remote nfs share?
rsynic command line your using...

file use case to not need no_root _squach

I don't realy know how to use pgadmin. I know that it a good alternative to having users run termianl commands. I had issues with working immich postgress and getting pgadmin to connect and make a backup..
sorry I can't help wiht pg admin.

The bes I can do is sent you to youtube and other guides...
https://www.youtube.com/watch?v=21J06XEbLD4

-This is because I can console into the database docker and run sql commands.

Regarding the another question running in my head
yes yo can remove old docker networkings you created.

docker network ls
docker netwrok rm <dockername>

see docker docs: https://docs.docker.com/reference/cli/docker/network/rm/

it will only cause issues if it is in use with another docker.

docker 0 is the interface made by teh docker/dockerd sytem application and is used for the docker bridge driver.

the br-53a22d28af07 is a shim bridge that conects over and uses teh br0 netork to give the docker a lan IP. This tells me that you are running a docker at a custom br0 and have givein it a static IP.
This is normal and assit docker and the container application at asigning a IP off the interface of br0

review more info here:

Review the 321 rule of thumb...

https://www.techtarget.com/searchdatabackup/definition/3-2-1-Backup-Strategy#:~:text=The%20basic%20concept%20of%20the,data%20is%20sent%20off%20site.

Most often do 221...

Ideally Clustering with a similar Hardware, same disk configuration and essential you duplicate the current unraid machine. then you rsync content between them.
Raid is not a backup it is redundancy and helps with maintaining uptime.

I make a share called Backup then use this folder on unraid on disk for a staging area. Once I ahve all my backup targets I have this one folder to send to another machine... I also use zfs and use snapshots and in a sftp ring I send my zfs snapshots as a backup...
https://docs.oracle.com/cd/E18752_01/html/819-5461/gbchx.html

I have a cron script... ZFS utilizes snapshots and the zfs send command for efficient backups and replication. Snapshots capture a point-in-time view of a dataset, and zfs send creates a stream of the snapshot's data, which can be sent to another location (another pool, a file, or another system) using zfs recv. This allows for incremental backups, sending only the changes between snapshots, significantly reducing backup time and storage space. 

....

I also implement and utilzes a few CA Apps and plugins.

Such as Appdata backup, zfs backup buddy, unraid replication. rclone...

There is alot of this talk and examples of script and commands you can do to automate and backup
example:
https://forums.unraid.net/topic/126028-simple-automated-offline-backup-script/
https://unraid.net/blog/unraid-server-to-server-backups-with-rsync-and-wireguard

and
https://forums.unraid.net/topic/97958-rsync-incremental-backup/
do yo have any specific question?
a backup is only good on how you handle your recovery plan and known it can be restored from the files you have backed up
The main tool I use that does just about everything to backup locally is the appdata backup plugin.
Unraid flash dirve, dockers, vm templates...
https://forums.unraid.net/topic/137710-plugin-appdatabackup/

some side Info That I have found over the years:

VMs
With VM snapshot this has gotten tricky...
https://forums.unraid.net/topic/169220-vm-snapshots/
Unraid libvirt image needs backup ed.
Snapshot Database needs backed up. Usually separate boots to clonezilla is needed to image a the entire vdsik and a copy of the vms xml code is copied to be resued for a full restore of a vdisk image and configurations...
Plugins exsit for vm backup...
I use app data plugin to grab the VM metdata and xml configs...
vm backup plugin is ok you really have to watch the use of vm snapshots and bring vms to 1 vdisk file...
https://www.youtube.com/watch?v=ntjQphOSPPI
https://www.youtube.com/watch?v=rDGSS1iEtq4
https://forums.unraid.net/topic/86303-vm-backup-plugin/

Dockers
app data backup plugin.
sometimes the docker vdisk needs backuped. (Only if docker paths and data is misconfigured. thus, docker data is saved within the docker image and not a volume set to a path to store on unraid...
snapshots of the appdata folder with zfs I used to send and backup... zfs snapshots are automated 1 a month and i trim them to have 12 so 1 year worth of backups.
with appdata I can configure what i'm backingup on the docker and have it tar the docker image contents and grab some volumes for it.
*-Not ideal for things like plex/jellyfin/embry etc... as it will try to backup the entire movie library and may need additional configuration

General file backup?
copy paste... unasnged disk cont to remote devcie.
rsync / cp (copy command) / zfs send .... etc etc...

I'm partial to runign ssh and using ssh as a sftp server.
I then run fiel zilla (docker or cleint) and conect to my machein and use samba to the remote and file zilla tranfer files.
This is becaue I can leverage samba to write as nobody and surs group for file access wiht the samb connetion but that a bit advace configuration...
-This also give me a change to scan a new file being written to storage for virus / malware checks.. aslo a bit advance setup...

Look into other plugins such as unraid replication. if using zfs zfs buddy. and teh vm backup plugin

My probe needs to wait until something else load which wont within to go file. (Too early in boot process)

You're better off using a user script plug-in and running the script at first to array start.

this seems more like a issue of vlan tagged and/or how you setup the vlans...

unraid server has a 4 port nic, each port is connected to a different vlan. (10 br0),(20 br1), (30 br5).

Usually unraid will show br0.10 to designate that the br0 interface over vlan10
*Vlan Networking on Unraid s is a pain to explain...

review post:
https://forums.unraid.net/topic/190501-unraid-in-unifi/#findComment-1558254
start at the top and go thorugh the whole post!

Unraid is not networking equipment... untaged / taged vlan on a interface issue...

############################

the br-#### is normal to see that is a software bridge containing over and made for the docker usually something set to a custom br0 - macvlan/ipvlan

Giving this info:


as the dockers don't see br0.10
unraid is using default vlan1 when calling br0 for this docker and ip range. Thus a cros talk miscommunication issue...

The fix is removed the br5, br1 and make a docker network at interface br5.vlan# etc etc...

This is called vlan tagging... Other unraid settings need to be checked and setup...

as example seen here
*You shouldn't have to disable bridging...

Make sure vlans are enabled in the unraid network interface tab...

under docker settings.

*I recommend using macvlan...
Enable host custom access
enable preserver custom docker settings...
Move to macvlan is set to ipvaln

*Macvlan is prefered due to isolation and the need for each service / device gettign a different mac address...

under vm setting make sure the vm netowrk is not set to vmbir set it to br0
as this changes things with unraid vm hyperv switch taht was a breaking point for how soem docker networks were created...

then proced to make docker networks at br#.vlan at the ip scheme... (YOU MAY NEED TO DELETE EXISTING DOCKER NETRWORKS!
Double check newotk settings . layer 3 swith settings . router setting to confirm a bridge...

use console and test dockers suh as docker run bash connected to your docker vlan...
Do to isseu with unraid and otehrs networking setups espcal with thoses who won't follow the instruction as seen in that post... I will no longer give commands nor help any further than saying it is possible but not easily done and setup on unraid due to many points of failure and netwroking hardware one uses...

I hate unraid with Vlan setups wish its was easier and done in the web ui... unraid is not networking equipment...

--May put in a feature request to have the ability to make docker networks within the web ui. and a area for vlan taging a interface... a lot of php code and terminal networking commands woudl need pre-programmed to make that work... doubt the devs will implement... (as I moved away form unraid do to similar issues and the inability to edit br0 docker network...

good luck...

install the docker:


set this to auto run at unraid reboot.

This will add more vm icons the doesn't stay running and will reinitiate the download.

Most likely at boot there was a network error and a inabilty to access githib for the system to download the VM icons.

Yes, your strategy is sound. You can use a test server to pre-encrypt drives and move them over. You can do this in full or in parts depending on how many drives you can afford to buy now. Just be very careful with the "New Config" step and assigning drives correctly afterward.

Encrypting the Array
You’re correct: Unraid requires encrypting a disk individually, then copying data onto it, round-robin style. There's no in-place encryption.
-Be sure to copy the encryption key for the array!!!

Full Migration Option (Test Server to Main):
Steps:
Set up test server with enough storage.
Encrypt all test server drives with your chosen format (xfs-encrypted or btrfs-encrypted).
Copy all data from main server to test server.
Backup the USB flash from the main server (important: use "flash backup" in the web UI or copy config/ folder).
Remove all old drives from the main server, including parity.
Move all test server drives to main server.
Boot up, choose Tools → New Config, assign drives manually.
Ensure nothing is formatted. Confirm all drives are recognized and contain correct encrypted filesystems.
Assign new parity, rebuild parity.
This will work, provided:
You don't format any drives accidentally.
You confirm encryption keys/credentials are correctly entered when the system boots.
You double-check disk assignments before starting parity build.

Partial Migration Option (Budget-Friendly):
Yes, this is a safer and more incremental option:
Copy only the data from the 11×4TB drives to encrypted drives in a test server.
Remove the 4TB drives and parity from the main server.
Replace them with new 14TB encrypted drives.
Use New Config, assign data + parity drives manually.
Build parity.
Round-robin encrypt the remaining drives (8TB/14TB) later, one by one.
This approach:
Lets you reuse 4TB drives for other tasks (or sell them).
Minimizes upfront cost.
Reduces risk by not touching existing encrypted drives.

Important Considerations
File System: Use xfs-encrypted unless you have a reason to prefer btrfs-encrypted. xfs is generally simpler and more stable for single-drive volumes.
Encryption Key: Store your encryption passphrase or keyfile securely (ideally outside Unraid).
New Config:
Use only when you're 100% confident about drive contents.
Always check "Parity is already valid" only if you know it is — in your case, it won’t be, since drives were moved/formatted.
UUID Conflicts: If reusing disks, always wipe old partition tables (wipefs, blkdiscard, etc.), or Unraid might get confused.
Unraid scripts exisit for encryption array and grab key files...

some things to review:

Data Encryption | Unraid Docs
Data Encryption | Unraid Docs
Unraid supports the use of encrypted drives in both the cache and the array. It does this using the Linux LUKS (Linux Unified Key System) encryption modules.
Medium
Automating Unraid Array Decryption
As an Unraid user, manually entering encryption passwords at every boot can become tedious and inconvenient, especially for your family.
A the the forum post on unraid encyptions:



I'm aware of the new encyption key and script but not sure of its setp and instalation...

Several scripts and methods exist for managing encryption in Unraid systems. One notable script is unraid-newenckey, which allows users to change the unlock key of their encrypted drives. This script ensures that the new key is backed up safely, as losing the new key can result in data loss.
https://github.com/doron1/unraid-newenckey

Gist
UnRaid New Encryption Key
GitHub Gist: instantly share code, notes, and snippets.

...

https://forums.unraid.net/topic/98978-plugin-nvidia-driver/

open source driver only noted in bottom right of settigns page for drivers for 5000 series cards...

sounds like database coruptions..
as it specfic to sqlite...
(More on that latter...) as that apears to be the core issue here...



To Inform:
Unraid runs at the root user level to gurantee file access. However, Linux OS files type permission are run at the no user(nobody) and set to group(users) permission for all users defined in users group

Each User and group has a ID # associated...

Unraid host
root 0 ID
nobody 99 ID
users 100 ID



Jelly fin may require additional data such as PUID and PGID

this said due to general file access and permission sometimes that is not enuoth and there is a need to run as the root user... the pgid and puid is set to 0





Unraid has a system tool in place to set "docker safe permission"
This essentials set chmod 777 and chown nobody:users to folders and files...

cd /path of jelly fin data

Docekr safe permission for Unraid

cd /mnt/user/appdata
chmod 777 -R *
chown nobody:users -R *

to fix file access and permission

The issue appears to stem from some form of Postgres database corruption. As long as the physical file exist this is recoverable.

Your Postgres appears to have broken and missing metadata and additional data for files that are either in a different path or moved when original metadata was grabbed / existed.
(Per the log)

As the admin account at first creation.
Click the profile at the top right:



and go to administrations:


at the tabs in the Left go to jobs:


Here we need to run jobs to collect and refetch data that existing and is missing...

I would recommend running all the jobs clicking missing next to each option in this.

Example:




These will scans and will put the unraid system under load. Jobs will need to finish before touching or editing Postgres database.
I don't remember which one as it may be a grouping of multiple depending on the errors. BUT, As long as the file itself exists, then its a matter of telling immich to rescan and refetch missing and bad data on a file.
Unriad CA uses image genius docker image (They do not acknowledge the vectorchord plugin yet and have updated the image accordingly...) Please stay and/or use the Postgres with pgvector to insure immch is happy via unraid CA...
Immich the Maintainers themselves, wants us to use there docker compose and using immich-app docker image for immich server along with the new vectorchord... (Compose file only.)
- Both will work with different Postgres database (due to immich change and its very confusing on which you are using...) immich-app immich server docker doesn't exist in the CA and will not due to compose settings incompatible with Unraid CA template deploy and docker management... (This is more to do with the 2 separate immich server and immich machine learning dockers in the compose file where image genious combines the 2 in 1 docker image...
Please Confirm If you were running CA with the tensor Postgres or the immich-app Postgres with vectorchord as I added immich-app Postgres database to unraid CA for Compose Immich users when database migration and updates occur.


See Docs:
https://immich.app/docs/administration/system-integrity/
https://immich.app/docs/FAQ#does-immich-support-reading-existing-face-tag-metadata

also the unriad immich docker CA template support thread is located here:
https://forums.unraid.net/topic/146106-immich-docker-self-hosted-google-photos-setup

You need to pick one use unraid CA Templates:
-Postgres Tensorchrod Database
-Immich server via imagegenious
in which you can fresh installation per this guide: https://forums.unraid.net/topic/190531-guide-immich-docker-setup-unraid-ca

Or you run immich server via immich maintainers CA. which you can freshly install per this guide: https://forums.unraid.net/topic/190532-guide-immich-docker-setup-docker-compose

You can move and maintain existing data as long as you map and keep your volumes correct. Due to the vector plugin change, I personally would recommend running the compose variant. Those still on and using the unriad CA template install should stay on and use the maintained with unraid CA template using data per imagegenious moving forward...
https://github.com/immich-app/immich/releases

Small docker compose update... The above compose will work, but if docker compose down, and docker compose up. This may make duplicates machines in the tailscale admin interface, This is due to the tailscald not having a state file and keeping machine identity... A simple command file to the tailscale sidecar docker method... as the machine identity is not saved and is removed at compose down removal...

Simple fix though.. Add the tailscal state file:
command: tailscaled --state=/var/lib/tailscale/tailscaled.state
*add this under tailscale docker compose section...

With this options added. I had to remove the machines form the tails scale admin interface and when I composed up I only had immich and plex no omv (the urnaid host)

Updated example Fix tailscale state file
#version: "3.8"
services:
tailscale:
image: tailscale/tailscale
pull_policy: always
# Watch Tower Compose Update Docker
# labels:
# - "com.centurylinklabs.watchtower.enable=true"
container_name: tailscale
network_mode: "host"
privileged: true
cap_add:
- NET_ADMIN
environment:
- TS_AUTHKEY=${TS_AUTHKEY}
#Get your auth key here: https://login.tailscale.com/admin/settings/keys
volumes:
- /dev/net/tun:/dev/net/tun
- /mnt/user/appdata/reverseproxy/tailscale/var-lib:/var/lib/tailscale
- /var/run/tailscale:/var/run/tailscale
command: tailscaled --state=/var/lib/tailscale/tailscaled.state
restart: unless-stopped
tsdproxy:
image: ghcr.io/almeidapaulopt/tsdproxy:latest
pull_policy: always
# Watch Tower Compose Update Docker
# labels:
# - "com.centurylinklabs.watchtower.enable=true"
container_name: tsdproxy
ports:
- "8080:8080"
#If you have something else using port 8080 change to 60080:8080
volumes:
- /var/run/docker.sock:/var/run/docker.sock:ro
- /var/run/tailscale:/var/run/tailscale
- /mnt/user/appdata/reverseproxy/tsdproxy/data:/data
- /mnt/user/appdata/reverseproxy/config:/config
restart: unless-stopped
#Optional Reverse Proxy
# npm:
# image: jc21/nginx-proxy-manager:latest
# pull_policy: always
# Watch Tower Compose Update Docker
# labels:
# - "com.centurylinklabs.watchtower.enable=true"
# container_name: npm
# hostname: proxy
# networks:
# netproxy:
# ipv4_address: 172.42.0.4
# ports:
# - "127.0.0.1:8180:80"
# - "127.0.0.1:8443:443"
# - "8181:81"
# environment:
# - DISABLE_IPV6=true
# volumes:
# - /mnt/user/appdata/reverseproxy/npm/data:/data
# - /mnt/user/appdata/reverseproxy/npm/letsencrypt:/etc/letsencrypt
# restart: unless-stopped
#Network for npm docker bridge and tsdproxy
networks:
netproxy:
driver: bridge
ipam:
config:
- subnet: 172.42.0.0/24


Recommend fixing to not cause tail scale device duplication:
*I also recommend running watch tower as host setting labels to update compose dockers on unraid...

What the admin interface will look like after comand added:


docekr-compose.yml


Watchtower docker in compose
#version: "3.8"
services:
watchtower:
image: containrrr/watchtower
container_name: watchtower
restart: unless-stopped
network_mode: host
volumes:
- /var/run/docker.sock:/var/run/docker.sock
command: >
--schedule "0 1 10 * *"
--cleanup
--include-stopped
--label-enable
labels:
- "com.centurylinklabs.watchtower.enable=false"

the erro your getting is that immich can't talk and conect to your postgress docker:

Error: connect EHOSTUNREACH 192.168.1.178:5433
at TCPConnectWrap.afterConnect [as oncomplete] (node:net:1634:16) {
errno: -113,
code: 'EHOSTUNREACH',
syscall: 'connect',
address: '192.168.1.178',
port: 5433
}

what is your postgres docker connection settings?

default Postgres port is 5432

?is Postgres running in what network mode?

is it running host mode (it will be at your unriads ip address) usiung port 5432

is it in bridge mode?
ther is a know tempate issue with teh continer port being a typo of 5433
in bridge mode it will still be unriad ip but you can move the port with the template setting...

custom br0/eth0/bopnd0 at a static ip of 192.168.1.178
?is unraid the 192.168.1.178


Please edit the Postgres docker and corect or gather the docker network type and port.
Then edit teh immich docker to correctly talk to the Postgres docker.

This is normal you have to edit the inline xml code or use the web ui to remove the OS isntall disk and/ or disk drive complete when the Virtual machine is off.



Is the USB connected to another VM?

IF so when teh VM went off it was connected and now apart from another VM thus It was not avialable when that VM started...
Use the plugin USB Manger to assist with USB devices..



this will give you more otion at when this VM starts have xyz USB and vise versa... This seems more like mutiple USB added to mutiple VMs and not known what the device is connected to...

The USB at the bottom are "etherail" they exist and are 100% saved to teh XML like the disk drive early if they are not avilalbe when off they are in use else where...

I would highly recommend starting over then. in which case you can keep the cache mirror setup for btrfs. While we could go over share and disk setting per the unraid doc there some inherit concerns and questions that need to be done and some fixes would not be easy wiht out chaing and invoking mover which I don't trust atm due to recent build pushes and potential data loss...

So, where your at now. We can use the unraid tools option and run a new disk configuration...

Please stop all dockers. go to setting docker off
please go to vm manger and make sure vms are off.
then on the main tab stop the array.

Next we go to tools
new config
we can keep the cace disk pool...
as the goal here is to remove the array
removing the disk 1, disk 2, parity, parit 2.

then at the main tab set array slots to 0
(you may need to unassign the disk from the array at the top)

make a zfs pool with that grouping of disks.
Then under pool devices. you will add a 4 slot disk

I don't know call it data. Come up with a name... as this will become the array moving forward...

Click data(the pool name you chose) and select filesystem type zfs and chose raidz1 (or raidz2 you can lose 2 disks and keep all the data similar to your current setup)
Doing this will set your disk up as a true nas raid array for redundancy in disk.
(this will make a raid 5 type array where you can lose 1 disk and still have all the data on all drives.) data is striped and mirrored across the multiple disks...

the reason for zfs in your case is due to disk access and its ability to increase disk read speed and increase overall data...

as in your current array setup your leaving 14 TB of space where you can potently get atleast 7 TB if not more on ZFS.

Review video over unraid zfs
https://www.youtube.com/watch?v=UR5RCItyCsw

RAID IS NOT A BACKUP ITS! Redundancy and uptime to keep system services running...

with that done. We can look at some add-on plugin as linked in the top post as the plugin adds additional functionality into unraid...
we can then make "datasets" which become shares and use datasets for the start of your file hierarchy

review for examples as this need to be though out and setup before data is compiled making location where xyz will be stored...
https://trash-guides.info/File-and-Folder-Structure/

With each dataset, we can leverage zfs for snapshots of the data for backup purposes..


Example OF my main screen and some datasets...

Then we can go over editing the shares tab and fixing the sahres to make sure primary is set to the corect disk example of my dataset file hiracry


then at each share and other unraid setting, we can leverage nfs/samba to share that off the machine...

I would advise setting the docker and vm system image on cache and potential install the swap plugin and making a 2 GB swap volume for a just in case Out of Memory...

then we can go over some side improvement script such as automatic backups, tweaks and user script settings to run automatic snapshots...


Otherwise we can skip running zfs. You keep the potential ram resource that would be used running zfs, you leave the 7TB if not more on the table and relie on other script to simulate data snapshot of data that would exisit in zfs... then we can try to leverage the uinrad share tab and look into mover to move the data off.

as you will be fighting the setting here between the sahres and ther data..

-as the easeit way would be to come up with a file hireachy first make the new folders where you wnat to put them and refactor your data to be in taht file dataset...

and then we may need to do some weird sym linking to fix which can become a mess of things... where the dataset in zfs and backups and snapshots give you more data and ease access to setting file hireachy early fixing where your placing data...

As example, then all the Docker data can then become a dataset and date for docker can be stored in that dataset...


as, because you have a

This will complicate how we would need to proceed moving the data off and correcting where the data needs to be saved on the disk...

in zfs all datasets would exist at /mnt/pollname/datasename

so I have a pool called vm-zfs and a dataset called docker

Unriad system host is called Docker...
root@Docker:~# cd /mnt
root@Docker:/mnt# ls
addons/ cache/ disks/ remotes/ rootshare/ user/ vm-zfs/
root@Docker:/mnt#

using unraids fuse system after making the file hierarchy uses the datasets and setting up the shares of where data exist...

root@Docker:/mnt# cd vm-zfs/Dockers/
root@Docker:/mnt/vm-zfs/Dockers# ls
Plex/ appdata/ immich/ local_ai/ netbox-docker/ rocketchat/
root@Docker:/mnt/vm-zfs/Dockers# cd /mnt/user/Dockers/
root@Docker:/mnt/user/Dockers# ls
Plex/ appdata/ immich/ local_ai/ netbox-docker/ rocketchat/
root@Docker:/mnt/user/Dockers#

the unraid fuse system for default pathing will be hard/sym link to the array for data...

then its a mater of updating docker setting and vm setting where on disk or fuse system is the data to be used...
as example plex temple pathing:

...

its better to make a new plex server on the new machine and use a plex claim to have 2 pelx servers.
Content search will still show up but will play from the other server.
 
I have done this a few different ways. Form a new claim ID and new/different plex server to unassigned disk remote access the other machines from a shared folder path (smb/nfs)
and adding another path for that folder to the docker and that folder path within the docker via plex library addition.

The reason i say its better to run a second server is how plex and unraid will handle when one server is offline compared to the other.

otherwise

New server
-new pool disk setup
a share is made and shared either via nfs or samba...
For this example, lets call that path /mnt/user/remoteplexdata
Make sub folders within this folder...
mkdir tv
mkdir movies
mkdir music
mkdir photos
 
 
Old server
-Unassigned disk plugin
*Connect to ne ser either via nfs/samba
you should now have a path on the old server that connect to 
/mnt/user/remoteplexdata
 
with unassigned disk plugin connect to remote share.
so now old server /mnt/remotes/pathname 
is linkesd to new server /mnt/user/remoteplexdata
 
cd into 
ls will show /mnt/remotes/pathname 
movies tv music photo...

^ if you see the folders on the new server Perfect, you are now ready for plex to use those paths...

############
edit plex docker
add a new variable path

container path lets use /remoteserver
map it to /mnt/remotes/pathname 
 
start/apply the docker
now open plex and go to libray and libar point add /remoteserver/movies for movies etc etc....

This accomplishes what you want, but only when both systems are working. a network outage, system reboot, unassigned remote connect can break due to communication issues this way. While this will work it is safe and less disk thrashing on the disk to run a second plex server on the new server... This way when a server issue happens and one of the systems are offline you still have part of your plex media running...