58 minutes ago, pille said:

Thanks for that very good and nice reply. I just thought that it might help with performance. But with your information I quess for me it is now a little bit better to dive into gatherings whole shows etc. instead of spreading them all around. Great info especially with the pictures. The only thing I might see with performance is, that all new stuff will be at the new drive, which is already a used one. And then there might be multiple streams to multiple clients, always from just this one drive. That does not really matter then? Do you also have a recommendation for the split level?

If it were me, I'd take a wait and see approach. If you happen to run into performance problems using defaults then you can take measures to try to alleviate those problems. Organizing different things onto specific drives can help in situations where you're running into throughput problems from a single drive, for sure. And the split level setting can help you out if you do end up having to worry about such things. But if you never actually run into the issue then it probably isn't worth your time to worry about it and try to do something about something that isn't actually happening for you.

I suppose there might be a case for just setting it up to avoid that situation from the get go, but I'm lazy. If it is working fine without me having to do anything, then that's good enough for me. If I run into problems then I'll do something about it. Everyone's different. Some like setting things up for the worst-case scenario straight away during initial setup. Maybe they've already got an idea of what the typical workloads will look like and they want to think ahead to avoid headaches down the road. I'm not sure if I'll ever run into performance bottlenecks, so I'm just waiting to see if it ever is an issue. If it never crops up, that's some good luck. If it does rear its head, I'll take some steps then.

Can I plugin a NTFS drive then use this to move files to the array?

2 hours ago, Sadaa said:

Can I plugin a NTFS drive then use this to move files to the array?

No, unbalanced is for transfers within the array. You use the Unassigned Devices plugin for that.

Edited December 18, 2025Dec 18 by Kilrah

Hello,

I'm trying to transfer a 4Tb drive into 3 2Tb drive.

It tells me that my 3.8 Tb folder cannot be transferred because it doesnt have space

Cant this plugin "dispatch" each file into different drives? Like if this file was in my cache the mover would divide it between the 3 disks

46 minutes ago, Olick said:

Hello,

I'm trying to transfer a 4Tb drive into 3 2Tb drive.

It tells me that my 3.8 Tb folder cannot be transferred because it doesnt have space

Cant this plugin "dispatch" each file into different drives? Like if this file was in my cache the mover would divide it between the 3 disks

Someone smarter than me can correct me but I think you’d have success if you drilled down into the folder and selected multiple smaller folders. Either all at once or multiple batches. I think selecting the main drive or main share sums up all the data as one file “size”. It will “dispatch” 3800 GB but it can’t fit 3.8 TB on a 2 TB drive in theory. I know that’s not what you’re trying to do, but that’s how the app sees the data. Maybe someone can explain this better than I can.

Well, Unbalanced broke 95 of my hardlinks that I had on sonarr. Guys don't forget the -H flag if you have a bunch of hardlinks, spent 2 hrs on Sonarr importing those shows back.

Edited December 23, 2025Dec 23 by Olick

A tiny issue I noticed in Unbalanced: the time elapsed and time remaining are formatted differently (at least in English): for elapsed there's a space between units, while for remaining there is not (see the attached screenshot). It's purely cosmetic. And to be really picky it would be more consistent to have leading zeros for minutes and seconds (so that the text doesn't jump around when going from 9 to 10, for example).

I'm new to unbalanced and I could be missing something, but I'm not seeing it. That is, I would have thought there was an option to basically select a number of drives and then have the app disperse the data evenly amongst those drives. What I'm seeing instead is that this is a very manual process where you need to specifically point at what data you want to go where...right? I'm not trying to "dink" the app, just trying to understand.

Also, I'm assuming this doesn't exist so I thought I would ask to be sure. It sure would be nice to have an option to select multiple folders at once (like a SHIFT - LEFT CLICK option) instead of manually having to click each and every folder you want to have moved. Again, maybe I'm missing something.

There's usually no reason to disperse across multiple drives, that's actually usually detrimental, so why do you want that?

That said I believe you can achieve it in with the appropraite share settings.

7 minutes ago, Kilrah said:

There's usually no reason to disperse across multiple drives, that's actually usually detrimental, so why do you want that?

That said I believe you can achieve it in with the appropraite share settings.

Just to balance data across the drives. I think the school of thought some time ago was that having something like 50% - 60% of data drives filled was sort of the sweet spot. If a drive got somewhere around 75% filled, then you needed to add more drive space. Evenly distributing that data out could help with that.

BUT now that you mention it, I think I read somewhere recently that is no longer the case. Drives being unbalanced isn't really the end of the world. So, I guess I shouldn't worry about it.

23 hours ago, JP said:

Just to balance data across the drives. I think the school of thought some time ago was that having something like 50% - 60% of data drives filled was sort of the sweet spot. If a drive got somewhere around 75% filled, then you needed to add more drive space. Evenly distributing that data out could help with that.

BUT now that you mention it, I think I read somewhere recently that is no longer the case. Drives being unbalanced isn't really the end of the world. So, I guess I shouldn't worry about it.

I do this too when I add new disks because I want to spread the I/O load across all disks instead of having the new disk be responsible for every new file until it's as full as the rest.

I use unbalanced to do it by first setting the minimum free space (unbalanced setting) to free space on the array in total divided by number of disks. Eg. 5TB free and 10 disks = 500GB minimum free.

Then I put together a quick Javascript macro to randomly select a bunch of files on disk 1 to transfer to the new disk. Rinse and repeat for all disks.

It's not super elegant, but it works.

I don't subscribe to the common philosophy around here of spinning down disks as much as possible. All of mine are always spinning and an equally random chance that any given file is on any particular disk gives the best chance that the file I need can be read from a disk that isn't busy, on average.

Hi,

I know this plugin can be turned off or on easily but when its running, everyone in the network can see all the file names in the entire server by going to [server ip]:7090.

Is there any way to put this plugin behind a password. Although no file content is exposed this seems to be a big issue.

17 hours ago, tTIKA said:

I know this plugin can be turned off or on easily but when its running, everyone in the network can see all the file names in the entire server by going to [server ip]:7090.

Is there any way to put this plugin behind a password. Although no file content is exposed this seems to be a big issue.

For an outside solution, in this post I explained how I put unbalanced behind a docker reverse proxy that requires authentication (traefik + authelia, in my case). It should possible for other reverse proxies and authentication providers to do the same thing. If you don't use a reverse proxy, you can still do the iptables step to limit who can access it locally, if your LAN is set up with sticky IP addresses.

Edited February 17Feb 17 by nearcatch

On 12/23/2025 at 2:28 AM, Olick said:

Well, Unbalanced broke 95 of my hardlinks that I had on sonarr. Guys don't forget the -H flag if you have a bunch of hardlinks, spent 2 hrs on Sonarr importing those shows back.

Just had the same. It would also prematurely end the transfer while there were still 90GB left.

TL;DR
Share must be cache primary [secondary doesn't matter] when doing top level manual splitting via rootshare

I don't know where to put this.
But just spent all day trying to figure this out with LLMs and finally fixed it myself because they kept thinking this was a minimum space or mounting issue (it isn't).

Moved my music from array to cache drive with unbalanced.
Found out I couldn't write to music (disk space full).
Spent entire day circling trying random commands.
Finally just set my Media share to cache first (as my other share was working).
Worked first try.

Dang.

released 2026.04.25 (whoa, its been more than a year since the last upgrade)

• added auth support (local, backed by a disk file)

this should improve security a bit, will potentially look into ssl certs down the road

pls report any issues you find

the happy path would be

• stop unbalanced

• enable auth

• start unbalanced

• on first access to the page, it will request you to set the password

• access from other browsers will ask for username/password

• you can logout from the unbalanced ui (top right)

released 2026.04.25a

• migrated encryption algorithm to argon (vs previous bcrypt)

this makes password length limitless, for practical purposes im maxing it out at 256

released 2026.05.02

• hardening of file-touching commands (no longer shelling out, guarding paths, use golang primitives)

• drive operations through plan ids rather than client supplied operation details, to prevent potential injection of unsafe paths or transfer instructions

this release is about improving the plugin's security envelope, i actually never heard of a security incident, still better safe than sorry

there's one or two additional security related changes, but i won't throw another update for the next week or so, in case any of any issues

Edited May 2May 2 by jbrodriguez

small update 2026.05.02a

• rsync flags hardening (rejects some flags that prevent normal unbalanced operation)

• small adjustment to replay logic (to avoid altering history)

released 2026.05.03

• improve login/logout user experience

• additional hardening of different surfaces (some shell invocations, websocket auth, rsync execution paths, .env permissions and logging)

wanted to land these changes, before the last (bigger) one in the security category

I'm having an issue emptying one of my 28TB drives on unbalanced v2026.05.03 using the scatter and move options. I have 15 total drives with 110TB of free space across my array. I have the unbalancer settings set to keep 2TB of free space on each drive, so there is still ample room for the 28TB. I check all the drives with more than 2TB of free space as targets to hold the 28TB I need to move, but unbalanced just fills up the first drive completely and then fails, instead of stopping at the minimum threshold set in unbalanced, and then moving onto the next drive. I can even see in the planning phase that it's going to attempt to put it all on a single drive instead of using multiple because of each disks free space bar graph not changing except for one, but it doesn't give me any errors or warnings. Is this a know issue or limitation, or do I just have something set up incorrectly?

Edited May 7May 7 by JohnnyBoyyo

9 hours ago, JohnnyBoyyo said:

... I check all the drives with more than 2TB of free space as targets to hold the 28TB I need to move, but unbalanced just fills up the first drive completely and then fails, instead of stopping at the minimum threshold set in unbalanced, and then moving onto the next drive ...

that would be a balanced move, but this is something built in into Unraid, with the allocationg logic (high water, etc)

many ppl here prefer to fill up disks, rather than having data scattered to serve as sort of "read-only/archives", there are some threads on the matter

what's unusual is what you mention about filling the first drive then failing, if you have the logs related to the error, i could take a look

finally, even though "balancing" disks is currently not supported, one of my goals is to have pluggable algorithms/logic, to move/copy files around: one of these algorithms could be to balance disks, versus the default one: fill one disk, then the next and so on.

not there yet tho, hopefully sometime this year

released 2026.05.17

• implements a time window to calculate speed (and remaining time)

this uses the last ~90 seconds to calculate speed (instead of the previous fixed 60 samples), should prevent some integer overflow that could have been happening

Edited May 17May 17 by jbrodriguez

I am trying Unbalanced, and like the purpose and how it works.

However, sometimes the planning phase (at least on my Unraid server) takes a very very long time, so long, that I'd rather cancel.

Is it possible to cancel the planning phase? And if not: is it difficult to create a cancel button on the planning screen?

Thanks!

ill add an issue to look into this