Hallo,

 

hätte eine Frage zu der High-Water Allocation Methode von Unraid. Ich mit mit meinem Server ursprünglich mit 2x18TB HDD gestartet (1 Array, 1 Disk) und habe sämtliche Daten von meiner alten NAS (4x6TB HDDs) kopiert. Danach habe ich die 4x6TB HDDs in den Unraid-Server eingebaut und dem Array zugewiesen.

 

Für mein "Multimedia" Share (Filme, Serien) habe ich die "High-Water Allocation" Methode gewählt sowie als Split-Level "Top-Level Directory". Wenn ich die "High-Water" Methode richtig verstanden habe dann sollte bei weniger als 3TB freiem Seicherplatz der größten HDD (18TB) auf die nächste HDD gewechselt werden. Trotzdem wird jedoch nach wie vor alles auf die 18TB geschrieben. Beim Share hab ich natürlich ausgewählt dass alle Disks verwendet werden sollen (auch einzeln auswählen ändert nichts).

 

mfg,

 

On 7/7/2024 at 1:48 PM, Neo78 said:

Beim Share hab ich natürlich ausgewählt dass alle Disks verwendet werden sollen (auch einzeln auswählen ändert nichts).

 

Hello!

 

I have similar question as the original thread maker. I have complete rebuilt my server and my disk setup is following 

Disk 1 14TB

Disk 2 10TB

Disk 3 10TB

Disk 4 10TB

 

And I choose under High water, automatically split any directory as required. But I have same problem when I copied my files to the server it start with the largest disk and then it was 30% full it went to the next disk and filled 30% and so on. I know a long time ago I could fill the disks simultanously I mean it writes to all the disks at the same time and filled them equally. Becuase I need fast access to my files as I have my family using plex to watch movies. 

 

Thanks in advance for any help in this matter.

 

12 minutes ago, andreas79 said:

Becuase I need fast access to my files as I have my family using plex to watch movies. 

 

When reading, every file is only on one disk.

Even if all Disks are running, the desired file is on one single disk and you can't read it faster as the single Disk is capeable to.

That has nothing to fo with high water oder so.

 

When you want to waste energy, you can switch of the automatic spin down and then every file on all the disks can be found/accessed without waiting for spinup. But even then: the file itself can only be transfered/read with the speed of that 1 Disk it resides on.

Edited July 11, 20241 yr by DataCollector

Hello,

this is Unraid and it don't use a typical Raid, which would increase disc write performance. Unraid uses a completly other method. To get fast write speeds you have to configure a fast chache drive, like a SSD or NVME. 

Hello!

Many thanks for you answers. From my understanding of "automatically split any directory as required" that I thought it was written a little bit of data to every disk. I understand that will spin the disks as it is used and that consumes more energy but that´s a trade of that I´m willing to take. As Archonw mentioned in that case a cache drive is needed. I will look into that.

Thanks again for all the help. 

48 minutes ago, andreas79 said:

From my understanding of "automatically split any directory as required" that I thought it was written a little bit of data to every disk. I understand that will spin the disks as it is used and that consumes more energy but that´s a trade of that I´m willing to take.

it would be then rather "split any file ..."

 

there is a reason this OS is named UNRaid ... you talk about a striped RaidX ... you can achieve something like this in a sep pool with zfs as sample ...

 

but this is all not this topic related