Hopefully this weekend I will get a chance to get everything set up on my first unraid server.  I want to make sure that I understand user shares and when to use them vs. standard folders.  My unraid server has eight 2TB data disks.  Here are my questions.

 

1.  Do most people only set up user shares for data that will (now or sometime in the future) exceed the storage capacity of a single disk?

2.  When you do set up user shares, do you limit it to the number of disks that you think you could ever possibly need?  For instance if I thought my movies share would never exceed 6 TB, should I assign my movies user share to 4 of my 2TB disks or all 8 of them?

3.  Aside from reducing the number of files lost if 2 disks simultaneously crashed, is there any advantage of setting up a user share vs. a windows folder for files that will only ever consume a small amount of the storage space on a single disk?  Let's say I have a Recipe folder that has 1,000 small text or word files that only uses a few GB's of space...is there any reason to set this up as a user share rather than having all this data reside on one disk under the Recipe folder?

 

thanks,

Murray

My user shares are:

Music

Movies

TV

Storage

 

Storage is my catch-all for things like your recipe folder or system image backups.  I know it's in storage\recipes\, I don't really need a user share for it.  Music, Movies and TV however, I want user shares for so I can point a media scraper at them and know it's only going to find a particular type of media.  I personally don't care what physical disk the data goes to, so I usually set each share to utilize all disks. 

Any top level directory you create on a disk will appear as a user share when you turn-on user shares. So, if you created 50 top level directories then you'd get 50 user shares.

 

User shares are created by taking the top level directories from all the disks and sharing them. The user share becomes a combination of the data on multiple disks when there are matching top level directory on multiple disks. For example, the user share "Movies" is the combined contents of the "Movies" directories on every disk. User shares are a neat way to combine data on multiple disks as well as allowing the user to create rules on how data is written to multiple disks.

 

So, you just pick which shares you want and if you want to limit their disk usage and go from there. As for your example, if you create Recipes as a top level directory then it becomes a user share. You could also create it as a subdirectory under a user share such as "Storage" or "Cooking" where other data is stored as well. You could also set this share to only use 1 disk if you wanted.

 

Peter

 

Hopefully this weekend I will get a chance to get everything set up on my first unraid server.  I want to make sure that I understand user shares and when to use them vs. standard folders.  My unraid server has eight 2TB data disks.  Here are my questions.

 

1.  Do most people only set up user shares for data that will (now or sometime in the future) exceed the storage capacity of a single disk?

2.  When you do set up user shares, do you limit it to the number of disks that you think you could ever possibly need?  For instance if I thought my movies share would never exceed 6 TB, should I assign my movies user share to 4 of my 2TB disks or all 8 of them?

3.  Aside from reducing the number of files lost if 2 disks simultaneously crashed, is there any advantage of setting up a user share vs. a windows folder for files that will only ever consume a small amount of the storage space on a single disk?  Let's say I have a Recipe folder that has 1,000 small text or word files that only uses a few GB's of space...is there any reason to set this up as a user share rather than having all this data reside on one disk under the Recipe folder?

 

thanks,

Murray

 

Murray

 

I use "user shares" to eliminate the often confusing process of remembering "where" on "which disk" is any particular file.

 

For instance, I have a share for Movies. My users (home users) need not know which disk the actual movie they want is on, just go to the movies share and there it is. Same with my TV share.

 

On the same token, I have one disk that is used for user data, and I have multiple shares for that data. This disk is segregated, I do not want my videos and personal data mingled (personal preference). If this disk ever fills up, another disk can be added, and space expanded, without needing to move files around, or my users needing to have several different mounts to find their data. They just mount /usersharename and they are off and running.

 

I have disabled disk shares on my system. My theory is that no user needs to know exactly where there data is. This makes things transparent to everyone. It also allows me to access rights accordingly. User1 cannot mount/delete data fro /User2.

 

I even have a user ID for my video playing device (WD TV Live). The video player has the ability to "manage" files. Since I believe that I do not want to "manage" files using my video player, it has "read only" access to the various video sections (/Movies, /TV, etc). Also, since the video player is intended for video, it does not have access to any personal user folders.

 

 

 

I find the users shares function to be very handy and flexible to my needs. Assigning access and access levels is keeping a level of sanity and manageability to my system, and keeps peace in my house. I am the "IS" person for my house, and if I have to manage that, I want it to be easy for me.

 

This is just my way of doing it, but you may find some use for you own needs in the examples I have given.

 

Bruce

Thanks Rich, Peter and Bruce.  I think that I have a better understanding of it now.  So if I didn't want to have to mount 50 user shares, then I would be better off putting most of what are currently top level folders under a "User Data" or something similar top level directory and maybe have a few other top level directories like Movies, TV Shows, Backups etc.?

 

Is it any more or less efficient to spread user shares out over all disks vs. just a few?  Currently I have ~1TB of data to move from my old traditional raid system to unraid, plus probably 2 or 3 TB in DVD's/Blue Rays that I want to rip to my server.  So with 16TB of available space and only about 4TB of data to fill it right now, should I:

1.  Just limit my array to 3 disks right now?

2.  Put all 8 into the array and spread the data evenly over all of them?

3.  Put all 8 of the in the array but only write data to 4 of them until I need the extra space?

4.  Put 7 in the array right now and go with option 2 or 3, keeping one as a spare so if I do lose a disk I don't have to rush out and buy another one?

 

Thanks,

Murray

Thanks Rich, Peter and Bruce.  I think that I have a better understanding of it now.  So if I didn't want to have to mount 50 user shares, then I would be better off putting most of what are currently top level folders under a "User Data" or something similar top level directory and maybe have a few other top level directories like Movies, TV Shows, Backups etc.?

 

Is it any more or less efficient to spread user shares out over all disks vs. just a few?  Currently I have ~1TB of data to move from my old traditional raid system to unraid, plus probably 2 or 3 TB in DVD's/Blue Rays that I want to rip to my server.  So with 16TB of available space and only about 4TB of data to fill it right now, should I:

1.  Just limit my array to 3 disks right now?

2.  Put all 8 into the array and spread the data evenly over all of them?

3.  Put all 8 of the in the array but only write data to 4 of them until I need the extra space?

4.  Put 7 in the array right now and go with option 2 or 3, keeping one as a spare so if I do lose a disk I don't have to rush out and buy another one?

 

Thanks,

Murray

 

This is personal preference. I keep a spare ready in case of disk failure. My spare is used as a cache drive.

Ultimately that's a matter of personal preference.  What I would do is:

 

1) Preclear all disks.

2) Limit array to minimum number of disks (4, I guess?) plus 1.

3) Store all other disks in ESD bags next to the server (cold spare approach).

4) Add disks one by one as storage requirements grow.

 

My reasoning for the above is that it would bother me to have several empty disks just using power for no purpose.  I would rather have the disks sitting outside the server collecting dust, but not using any power.  However, preclearing all of them is important because otherwise your server would have hours of downtime every time you added a new disk.

 

I'm also assuming that you are using a server with hot swap drive bays or other trivial means of installing a new disk.  If it is actually a hassle to install a new disk, then that's a good reason to install them all right away.

 

Another option is to install all the disks, but only hook up power to the ones needed.

 

A third option is to use one of the disks as a warm spare, but that's only after you evaluate the pros and cons of using a cache drive.  This thread spells it out.

User shares are great for users like my wife as well as my XBMC based HTPC. She clicks on the network and sees the following Names, but I included "on Disk" for explination in this post. Primarly setup as read only except for Photos and Downloads since she or XBMC doesn't need to write to Movies or TV.

 

Movies "on Disk1&3"

TV on "Disk2"

Photos "on Disk4"

MP3's "on Disk4"

Downloads "on Disk4"

 

I then use disk shares \\tower\disk1-4\ to write to most of the time hidden so others on the network get any wild ideas.

 

I do as Rahajal and have a couple of drives precleared sitting in their bags as spares and as needed for expansion.

Thanks Rich, Peter and Bruce.  I think that I have a better understanding of it now.  So if I didn't want to have to mount 50 user shares, then I would be better off putting most of what are currently top level folders under a "User Data" or something similar top level directory and maybe have a few other top level directories like Movies, TV Shows, Backups etc.?

 

Is it any more or less efficient to spread user shares out over all disks vs. just a few?  Currently I have ~1TB of data to move from my old traditional raid system to unraid, plus probably 2 or 3 TB in DVD's/Blue Rays that I want to rip to my server.  So with 16TB of available space and only about 4TB of data to fill it right now, should I:

1.  Just limit my array to 3 disks right now?

2.  Put all 8 into the array and spread the data evenly over all of them?

3.  Put all 8 of the in the array but only write data to 4 of them until I need the extra space?

4.  Put 7 in the array right now and go with option 2 or 3, keeping one as a spare so if I do lose a disk I don't have to rush out and buy another one?

 

Thanks,

Murray

 

I also keep a spare, preclear'ed drive on hand for fast replacement. In my field, we always say, hardware will eventually fail (and software will eventually work, but that is another topic). I want to be ready at anytime to start the fixing process.

 

I almost always like Rajahal's suggestions, and have great respect for them. I too would not spin any disk that was not needed/being used. Put in enough to be used, with room for some growth, and save the rest outside the array, waiting for the inevitable.

 

At least that would be my answer.

 

Bruce

You can limit small shares (not too much data) to a single disk. Try reading this section on user shares, maybe it will help understand the different ways a user share can be filled (the allocation methods). Take note of the split level too because that might come back to bite you later.

 

http://lime-technology.com/wiki/index.php?title=Un-Official_UnRAID_Manual#User_shares_2

 

Also, you can write to the disk shares forcing the data to go where you want and then just read from the user shares. That way, you fill the disks how you want them filled.

 

Peter

Thanks everyone for your explanations.  I do have icydock 5-in-3's so adding/removing disks are no problems.  So I will preclear them all which will give me an added benefit of knowing right now whether the disk is good or not and just add enough disks for the data I have now.  Then add disks in as needed.  I guess I shouldn't have gone wild buying disks every time there was a good deal on them! 

 

FYI, the general recommendation is to run at least two passes of preclear on each disk.  Takes longer, but gives you better piece of mind.

what do you do for TV shows?

 

I'd like to be able to point sickbeard or plex at a single source and not worry what drive its on. but if there is a show that is ongoing, its possible I'll end up with a second level folder with the same name. From the wiki, that sounds like it'd create a conflict and the lowest numbered drive would win. How do you get around that?

 

 

eg

 

disk 1

- TV Shows

  - house

      - season 1

          s1e1.avi

          s1e2.avi

 

disk 2

- TV Shows

    - house

      - season 2

        s2e1.avi

        s2e2.avi

 

 

In this case, wouldn't 'TV shows' be a top level folder and therefore a user share (good, want this). But 'House' would conflict so only season one would show up as its on the lowest numbered disk.

 

Is this only an issue if I manually transfer files?

 

eg I have 6 seasons of house, currently split across two USB drives. If I simply copy each USB drive to an equivalent in unRAID, then setup user shares I'll get a conflict. But if I set up the user shares *first*, just with 'TV shows', and *then* transfer my TV shows folders from each USB drive, then unRAID will allocate them based on the split levels/allocation etc and there will be no conflict? Surely they end up physically on the disk in the same location as my illustration?

 

 

what do you do for TV shows?

 

I'd like to be able to point sickbeard or plex at a single source and not worry what drive its on. but if there is a show that is ongoing, its possible I'll end up with a second level folder with the same name. From the wiki, that sounds like it'd create a conflict and the lowest numbered drive would win. How do you get around that?

 

 

eg

 

disk 1

- TV Shows

   - house

      - season 1

          s1e1.avi

          s1e2.avi

 

disk 2

- TV Shows

    - house

      - season 2

         s2e1.avi

         s2e2.avi

 

 

In this case, wouldn't 'TV shows' be a top level folder and therefore a user share (good, want this). But 'House' would conflict so only season one would show up as its on the lowest numbered disk.

 

Is this only an issue if I manually transfer files?

 

eg I have 6 seasons of house, currently split across two USB drives. If I simply copy each USB drive to an equivalent in unRAID, then setup user shares I'll get a conflict. But if I set up the user shares *first*, just with 'TV shows', and *then* transfer my TV shows folders from each USB drive, then unRAID will allocate them based on the split levels/allocation etc and there will be no conflict? Surely they end up physically on the disk in the same location as my illustration?

 

 

 

There will be no conflict. This is the beauty of user shares. TV shows will contain a single House directory  that contains season 1 and season 2. Conflicts only occur when full path and file names match.

ah, didn't understand that fully in the wiki - thanks.

 

so I can just dump the entire root of each USB drive into a user share, and unRAID will distribute accordingly? great.

 

and I'm good to start with the free version - 1x2TB parity drive, 2x2TB data drives, check its all working, then buy the license and add another couple of drives, and the system will adapt?