I already asked a similar question about whether using Drivepool for the media files was a good idea, but it brought up a new one.
I’m running Plex on a re-purposed gaming computer, which has run out of space. I set up a Mediasonic ProBox HF2-SU3S2 4 Bay 3.5" SATA HDD Enclosure, connected directly to the server box by eSATA. In the computer, my Plex media files are on a 3TB drive that’s full (J:), while Plex, and the Plex Data Folder is staying on a smaller C: drive. The extra storage will come from adding a 4TB drive that will go in the new Mediasonic ProBox (more added later as needed).
After getting it all running, and installed Drivepool, much to my embarrassment, I realized that I didn’t know how to set up Drivepool for this use. I now have my old Plex media drive (J:), and the new 4TB drive (K:), making the new virtual drive DrivePool (L:). After reading the instructions over at the Stablebit site, I’m still a bit confused on the best way to set this all up for Plex media use. I thought that what I needed to do was move the media folders from the old J drive to the new pooled virtual drive L, then tell Plex to look on L for the media files. Is that all there is to it? After that is done, what have other users here of Drivepool found was the best way to set up file protection and backups.
That sounds like it should be correct (Not being a Drivepool user myself, I eventually went to Storage Spaces after MS ditched Drive Extender). I imagine once you’ve moved everything from J: to the drive pool, you could then add the disk that was J: into the Drive Pool so you could continue using the space.
However, I would make a suggestion. Instead of pointing Plex directly at a drive letter, I would put your media in whatever folders, and make those folders network shares. You’ll have to do this anyway if your PMS and file server aren’t the same machine. If they are the same machine, access the shares through \localhost and you won’t really lose any performance, and by abstracting the media location from the drive letter, you can shuffle things around underneath without having to change anything from Plex’s point of view. I suppose you can do all that through mapping and junctions, but a share seems easier to me.
Thanks MasterChiefmas for the advice. I tried the keep it simple techniques first, so today I moved all my media folder files to the new drivepool drive, told Plex where to look, and did a rebuild of the database. It works great, no problems. I do still have to figure out a plan for backups, and how to use DrivePool built in duplication.
Whatever you do do NOT place the “Place where local application data is stored” in the pooled structure. If the drive is local to the server in can be on the pooled drive but external to the pool itself. Drivepool allows this. That is you can have files on the drive in the pool and others that are not.
If you want to have the Plex data directory on your new drive then all you need to do is access the drive via its actual drive letter (not the pool’s letter) and create a data directory that is not in the pool directory. The Pool directory, where all the pool’s files actually are, is a long, apparently cryptic, name and is is also a hidden directory so it is easy to avoid.
I found that it is best for me to use the original 1 tb drive in my server for Plex’s entire system and use my pooled drives, of which I have 10, for media storage.
If you have any questions about DrivePool itself I have found that they are pretty responsive on their forums and they respond pretty quickly to questions.
As far as exactly how to set it up I suggest the following:
- Set up your pool just like you would for and data use.
- Do not use the pool for the Plex data directory.
- Create a directory in the pool for Movies and another for TV shows and, if needed, other directories for other types of files.
- Place your media in the pool and use folder duplication (X2) for the media folders you cannot afford to loose.
Remember the Pool drive should be treated exactly like a regular drive. As far as access goes there is nothing special about accessing data in a pool.
To reiterate the pool drive is simply one big hard drive and is used/accessed exactly the same as any other drive.
I do not really know why the pool itself should not hold the Plex data directory but in should not. In fact it causes nothing but problems if the pool is used for that and Stablebit itself says to not use the pool for that purpose.
I have been running Drivepool to hold my Plex videos/audios for a good long time and I have had zero problems related to the pooling. Drivepool is one of the best software investments I have ever made.
Thanks for the help guys. It’s working great
@Elijah_Baley said:
4. Place your media in the pool and use folder duplication (X2) for the media folders you cannot afford to loose.
@Elijah_Baley ,
So, if you have 2 copies of a movie then which one will be played?
@NewPlaza said:
@Elijah_Baley said:
4. Place your media in the pool and use folder duplication (X2) for the media folders you cannot afford to loose.@Elijah_Baley ,
So, if you have 2 copies of a movie then which one will be played?
The system only sees the first copy.
Replying for Elijah… It will only be seen as a single file. The 2 copies are for resiliency, much like having mirrored drives.
Or if you are asking about the technical aspect… DrivePool will read from the most available physical drive.
DrivePool manages that. The operating system only sees one on the DrivePool drive when accessing it and therefore Plex only sees one.
There is an easy way to get to both copies at the operating system level BUT the real directories that hold the parts of the pool that have been allocated to hidden directories that are named like:
“PoolPart_a_whole_bunch_of_numbers”
You never need to actually access that directory unless something really bad happens.
BTW: If you are creating a new pool and you are using a drive with existing data that you want included in the pool all you need to do is move the folders containing that data directly into the PoolPart directory on that drive. It will then quickly appear in the pool and DrivePool will, over time, move the data around creating duplicates and balancing the load over the pools disks as needed. You can use the pool while all that is happening because DrivePool already owns the files and never “moves” the file while another process has it open. But DrivePool “may” copy a file that is being used, like being played, because that is not intrusive at all.
Also DrivePool never touches files that preexist on a drive that is being placed in a pool. It just creates the PoolPart directory so your existing data is on the disk but NOT on the drive that DrivePool makes out of the directories that it creates on all the drives in the pool unless and until you move it into the pool.
@wxman3 said:
@NewPlaza said:
@Elijah_Baley said:
4. Place your media in the pool and use folder duplication (X2) for the media folders you cannot afford to loose.@Elijah_Baley ,
So, if you have 2 copies of a movie then which one will be played?The system only sees the first copy.
@sGarver said:
Replying for Elijah… It will only be seen as a single file. The 2 copies are for resiliency, much like having mirrored drives.Or if you are asking about the technical aspect… DrivePool will read from the most available physical drive.
These two replies do not match up. Does it read the first file or will it pick it at random
I’m not trying the be an a$$… I was thinking about getting it and thought with this ‘duplication’ thingy it could help reduce the resources on one drive if its on(and accessable from DrivePool) another drive to access.
@NewPlaza said:
@wxman3 said:
@NewPlaza said:
@Elijah_Baley said:
4. Place your media in the pool and use folder duplication (X2) for the media folders you cannot afford to loose.@Elijah_Baley ,
So, if you have 2 copies of a movie then which one will be played?The system only sees the first copy.
@sGarver said:
Replying for Elijah… It will only be seen as a single file. The 2 copies are for resiliency, much like having mirrored drives.Or if you are asking about the technical aspect… DrivePool will read from the most available physical drive.
These two replies do not match up. Does it read the first file or will it pick it at random
DrivePool just provides the one that it flags as the primary. It has nothing to do with “first” nor does it really know about “most available.” However it does know about speed.
If you have a cloud drive (a practice I do not recommend for 99% of the DrivePool installations) included in a Pool and a file resides both on the cloud drive and a local drive then DrivePool will provide the one from the local drive.
EDIT: DrivePool can easily be setup to handle “load balancing” and since it will assure that files are never duplicated on the same drive you do not really need to be concerned.
There is no actual “first” file. There are 2 copies of a file, one on 2 separate physical drives, presented to the OS as only 1 file via the pool drive letter.
As for which of the physical drives actually are read when a file is requested… this is based on which drive is idle at that moment, and/or least under demand.
DrivePool just provides the one that it flags as the primary. It has nothing to do with “first” nor does it really know about “most available.” However it does know about speed.
Yet in testing, I have watched DrivePool consistently read from the disk that I was NOT writing to and/or reading from with another file. So this response is based upon my direct observations.
@sGarver said:
DrivePool just provides the one that it flags as the primary. It has nothing to do with “first” nor does it really know about “most available.” However it does know about speed.
Yet in testing, I have watched DrivePool consistently read from the disk that I was NOT writing to and/or reading from with another file. So this response is based upon my direct observations.
DrivePool can easily change the file it considers as “primary.” It is quite smart at doing that and it will watch the drives for activity changing the flag as needed. That behavior looks exactly like what you are describing and it has gotten a LOT smarter over the last year or two.
I think we may be looking at the same behavior just a little differently.
They say there is more than one way to skin a mongoose and they all result in the same thing, “a skinned mongoose.” This mongoose gets skinned quite quickly and no matter what way the skin is produced it is still a mongoose skin and used in exactly the same way.
So, with this ‘duplication’ thingy. Could it could help reduce the resources on one drive if its on(and accessable from DrivePool) another drive to access.
Example;
Drive#1 -> Moives
Drive#2 -> blank. This drive is empty.
If you put both drives in the pool with duplication on drive1. Will it copy all movies from drive1 to drive2?
Drive#1 -> Moives
Drive#2 -> Copy of drive1.
These two drive have the exact same movie(drive 2 is a clone of drive1 if you will).
If you put both drives in the pool with duplication what will happen?
@NewPlaza You are looking at this almost correctly…
DrivePool is setup to use duplication (meaning create 2 identical files on 2 separate physical drives). Note that this means ANY 2 physical drives you have in the pool (ie drive 1 and drive 3 -or- drive 2 and drive 7). But the OS only sees 1 file via the pool drive letter.
So you setup a pool drive using 9 drives, turn on duplication, and EACH of your files will be copied to any 2 of the 9 drives AS Drivepool determines best for balancing. Thus your first file may copy to drives 1 and 2, and your second file may end up on drives 4 and 7.
EXAMPLE of multiple reads:
GIVEN: The file my mother is streaming happens to reside on drives 1 and 3.
GIVEN: The file I want to start watching happens to be on drives 1 and 4.
GIVEN: Drive pool is reading from drive 1 for my mother’s file.
RESULT: When I start to watch my movie, Drivepool will read it from drive 4 because it knows that drive 1 is already busy.
Note: there are many setting and these behaviors can be adjusted, but for simplification here we are ignoring this.
@NewPlaza said:
So, with this ‘duplication’ thingy. Could it could help reduce the resources on one drive if its on(and accessable from DrivePool) another drive to access.
Example;
Drive#1 -> Moives
Drive#2 -> blank. This drive is empty.
If you put both drives in the pool with duplication on drive1. Will it copy all movies from drive1 to drive2?
Drive#1 -> Moives
Drive#2 -> Copy of drive1.
These two drive have the exact same movie(drive 2 is a clone of drive1 if you will).
If you put both drives in the pool with duplication what will happen?
If DrivePool only has two drives to work with and you place data directly in in a sub directory of the PoolPart directory on one of the drives and you have duplication turned on for that sub directory the it will create and copy (duplicate) of that file in a directory named the PoolPart directory on the other drive. I will try to detail what happens with DrivePool. Remember all the PoolPart directories are hidden directories.
We will start with two drives and no pool:
“D:” has a directory called Dir1 containing files F1 and F2
“E:” is empty.
You then tell DrivePool to create a pool by combining D: and E:
“D:” still has a directory called Dir1 containing files F1 and F2
and it also has a directory called PoolPart-A_bunch_of numbers
“E” still has no root files but contains a directory called PoolPart-Different numbers
There is also in your operating system a new drive “F:” that holds currently no data and has the capacity of the free space from drives D: and E:
Then you move Dir1 from the root of D: into the PoolPart directory on the same drive
The result is that initially:
“D:” has an empty root directory except for its PoolPart directory and that directory contains the Dir1 which in its turn contains F1 and F2
“E:” has an empty root and an empty PoolPart directory
“F:” now contains Dir1 and the files F1 and F2 and those files can be accessed and used just like a “normal” drive.
Now you turn on duplication for “F:\Dir1” (X2) in the DrivePool application.
Drive pool immediately starts copping the “Dir1” directory from the PoolPart directory on D: into the PoolPart directory on E:
During all that F: only shows a single instance of Dir1.
That is you never see any second copy of Dir1 when accessing the pool.
You never should actually access the PoolPart directory except for initial moving of stuff into the pool. Once the pool is created as above you would copy/move new data onto F:. You should use your pool F: just like a regular Windows drive and, in general, never even look at the actual physical disks involved except to access data that you do not want in the pool. Remember that the only thing DrivePool knows about data that is not in a PoolPart directory is how much space it uses and that quantity is deducted from the total available space on the pool. Also keep in mind that DrivePool is best at managing its own space so do not try to micro manage it.
Whew! That was pretty long. I hope it was clearer than it felt as I was writing it.
EDIT: Duplication is at the folder or file level but NOT at the drive level as DrivePool does not allow management at the physical disk level.
@sGarver and @Elijah_Baley .
I now have a better understanding on how DrivePool works.
Thanks .
I have been struggling with setting up Drivepool and trying to figure out what the advantages are. This made it EXTREMELY clear and was super helpful.
