I was wondering on some guidance on overall drive capacity and what is a safe percentage? I have been for years moving stuff around to keep each drive at the same basically capacity level and I realized I should just take the drive out of rotation once its hit a limit, might be easier than moving data back and forth. If this isn't a good idea then I wont but if it is, whats a good percentage?

 

I have 4TB drives (end goal, 3 are upgrade-able still from 2TBs); 8 of them at the moment and about 800GB free. Would performance be effected or anything else if I left say 10GB free on each drive? Should it be more? Is less fine? I just know if I keep a drive that low in rotation i run into space issues since dockers don't understand to not use that drive.

I was wondering on some guidance on overall drive capacity and what is a safe percentage? I have been for years moving stuff around to keep each drive at the same basically capacity level and I realized I should just take the drive out of rotation once its hit a limit, might be easier than moving data back and forth. If this isn't a good idea then I wont but if it is, whats a good percentage?

 

I have 4TB drives (end goal, 3 are upgrade-able still from 2TBs); 8 of them at the moment and about 800GB free. Would performance be effected or anything else if I left say 10GB free on each drive? Should it be more? Is less fine? I just know if I keep a drive that low in rotation i run into space issues since dockers don't understand to not use that drive.

Don't understand what you are trying to accomplish. You should just refer to everything as user shares and let unRAID worry about what drive anything is on. Set minimum free on each user share to be larger than the largest single file you will write to that share and everything else should take care of itself.

Agree -- just let UnRAID handle it.

 

Technically there's no reason you can't simply fill a drive up entirely.  On my media server (mostly static content), 5 of the 14 drives have less than 1GB of free space.    Clearly they're never written to anymore -- and once they got down to a few GB I selectively saved the last few DVDs to them so take max advantage of the space.  There's no read performance issue with completely full drives.  There is a write performance "hit" with very full drives that use Reiser; much less so with XFS.

 

I was wondering on some guidance on overall drive capacity and what is a safe percentage? I have been for years moving stuff around to keep each drive at the same basically capacity level and I realized I should just take the drive out of rotation once its hit a limit, might be easier than moving data back and forth. If this isn't a good idea then I wont but if it is, whats a good percentage?

 

I have 4TB drives (end goal, 3 are upgrade-able still from 2TBs); 8 of them at the moment and about 800GB free. Would performance be effected or anything else if I left say 10GB free on each drive? Should it be more? Is less fine? I just know if I keep a drive that low in rotation i run into space issues since dockers don't understand to not use that drive.

Don't understand what you are trying to accomplish. You should just refer to everything as user shares and let unRAID worry about what drive anything is on. Set minimum free on each user share to be larger than the largest single file you will write to that share and everything else should take care of itself.

 

Honestly I didn't know that setting existed haha. So thank you! That was what was frustrating me, because without it things would never copy over, and it would backlog etc etc. THANKS!

I was wondering on some guidance on overall drive capacity and what is a safe percentage? I have been for years moving stuff around to keep each drive at the same basically capacity level and I realized I should just take the drive out of rotation once its hit a limit, might be easier than moving data back and forth. If this isn't a good idea then I wont but if it is, whats a good percentage?

 

I have 4TB drives (end goal, 3 are upgrade-able still from 2TBs); 8 of them at the moment and about 800GB free. Would performance be effected or anything else if I left say 10GB free on each drive? Should it be more? Is less fine? I just know if I keep a drive that low in rotation i run into space issues since dockers don't understand to not use that drive.

Don't understand what you are trying to accomplish. You should just refer to everything as user shares and let unRAID worry about what drive anything is on. Set minimum free on each user share to be larger than the largest single file you will write to that share and everything else should take care of itself.

 

Honestly I didn't know that setting existed haha. So thank you! That was what was frustrating me, because without it things would never copy over, and it would backlog etc etc. THANKS!

Turn on Help in the webUI. 

Lots of things are explained for each page.

I was wondering on some guidance on overall drive capacity and what is a safe percentage? I have been for years moving stuff around to keep each drive at the same basically capacity level and I realized I should just take the drive out of rotation once its hit a limit, might be easier than moving data back and forth. If this isn't a good idea then I wont but if it is, whats a good percentage?

 

I have 4TB drives (end goal, 3 are upgrade-able still from 2TBs); 8 of them at the moment and about 800GB free. Would performance be effected or anything else if I left say 10GB free on each drive? Should it be more? Is less fine? I just know if I keep a drive that low in rotation i run into space issues since dockers don't understand to not use that drive.

Don't understand what you are trying to accomplish. You should just refer to everything as user shares and let unRAID worry about what drive anything is on. Set minimum free on each user share to be larger than the largest single file you will write to that share and everything else should take care of itself.

 

Honestly I didn't know that setting existed haha. So thank you! That was what was frustrating me, because without it things would never copy over, and it would backlog etc etc. THANKS!

Turn on Help in the webUI. 

Lots of things are explained for each page.

 

its on, Ive just been here to long, before help. I set all that stuff up in 4.0.

Agree -- just let UnRAID handle it.

 

Technically there's no reason you can't simply fill a drive up entirely.  On my media server (mostly static content), 5 of the 14 drives have less than 1GB of free space.    Clearly they're never written to anymore -- and once they got down to a few GB I selectively saved the last few DVDs to them so take max advantage of the space.  There's no read performance issue with completely full drives.  There is a write performance "hit" with very full drives that use Reiser; much less so with XFS.

 

So its not working, maybe I am doing something wrong?

 

On the share I put a "Minimum Free Space" to 5GB for that specific share. My docker is writing to that share specifically and it placed a 3GB file (or tried); when only 1GB was left. Am I not understanding this correctly?

 

Basically what I want is to not have to move data around because a drive fills up. In its current state however, Sonarr still writes to that location regardless...

When it begins to write a file, it doesn't know how large the file will be. You set Minimum Free to be larger than the largest file you ever expect to write.

 

Just for clarification, here is how it is supposed to work:

 

Say you set Minimum Free to 5GB, and there is 6GB left. You write a 4GB file. Since there is more than 5GB left, it writes the 4GB file, then there will be only 2GB left. If you try to write another file, it won't write it there because the space remaining is less than the Mininum Free.

 

 

The other thing that will affect this is Split Level. Split level has precedence over Minimum Free, so if Split Level says the file belongs with other files it will write them where the other files are.

Are you sure the docker mappings and paths aren't restricting the write to a specific drive? Is the docker writing to the array or the cache?

When it begins to write a file, it doesn't know how large the file will be. You set Minimum Free to be larger than the largest file you ever expect to write.

 

Just for clarification, here is how it is supposed to work:

 

Say you set Minimum Free to 5GB, and there is 6GB left. You write a 4GB file. Since there is more than 5GB left, it writes the 4GB file, then there will be only 2GB left. If you try to write another file, it won't write it there because the space remaining is less than the Mininum Free.

 

 

The other thing that will affect this is Split Level. Split level has precedence over Minimum Free, so if Split Level says the file belongs with other files it will write them where the other files are.

 

My real world example:

 

Minimum set: 5GB. 2GB was free. It tried writing a video file to the drive that was 2.2GB. It failed and left a .partial~ file (I assume this is a Sonarr thing). The split level I have set on my TV Series share is "Automatically split only the top level directory as required."

 

My directory structure is \TV Series\The Flash (2014)\Season 03\file.ext

 

When trying to use that directory it failed.

 

So, bringing it all back I guess my questions are now:

 

1. Is my structuring bad? If so what would be recommended? The way it appears is Shows can be split per drive without issue.

2. If the split level is an issue on the Minimum split can drives just be excluded from the share but still have their data shown in it? That way I can manually fill the drive and "turn them off" from being written too but still show in the share.

3. If 1 & 2 don't solve it, I guess I am back to splitting up the space evenly and leaving this completely manual 

Are you sure the docker mappings and paths aren't restricting the write to a specific drive? Is the docker writing to the array or the cache?

 

So NZBGet writes to the cache. Sonarr takes from the cache and moves it to a volume on the array such as: /tv = /mnt/user/TV Series/

Reread the split level help VERY carefully. If you allow only the top level to split, your shows can't get spread as needed.

Reread the split level help VERY carefully. If you allow only the top level to split, your shows can't get spread as needed.

 

But they are, I have The Flash on numerous drives without issue. Is it because it should actually be on the Season folder and not the show? and if that is the case I guess I would want 2?

 

\SHARENAME\Show Name\Season ##\file.ext

\TV Series\The Flash (2014)\Season 03\file.ext

 

\\SHARENAME\level 1\level 2\file.ext

 

this way multiple seasons can be on multiple disks?

With that split level, if you try to write a TV Series file into the folder "The Flash (2014)" then it will write it onto the same drive where the folder "The Flash (2014)" already exists. That way it keeps the 2014 episodes for the series together so there will be no spinup delay when going from one episode to the next.

 

Share settings only affect writes to the share, not reads. So if you exclude a drive from a share it will not write files for that share on that drive, but it will still read files for that share from that drive. It is also possible to exclude a drive from all user shares in Global Share Settings, in which case the drive wouldn't be used for writing or reading any user share.

 

 

Reread the split level help VERY carefully. If you allow only the top level to split, your shows can't get spread as needed.

 

But they are, I have The Flash on numerous drives without issue. Is it because it should actually be on the Season folder and not the show?

Split level doesn't effect reading, only writing new files.

Split level doesn't effect reading, only writing new files.

 

I know, but having it split at level 1 the show is still being split across different drives... Should that not have been the case?

 

Ideally, with my scenario, am I to assume switching it to level 2 should fix my issue anyways?