setup disk

Hi.

Can it join 4 250 GB disk or 1 1 TB disk ??

I wish to use it as a server for my Media center 2005.

It has to store all my DVDs and music

Each data disk has it's own file system and can be accessed as a separate shares under My Network Places.  They show up as:

//tower/disk1

//tower/disk2

...

(where 'tower' is the computer name, which of course can be changed).

There is also another way of exporting a "composite" view of all the disk top-level directories as individual shares.  For example, if each of your data disk has the top-level directory 'Music',  then you can export a share called 'Music' which present a merged list of all the files under the 'Music' directory on all the data disks which have a 'Music' directory.

This is similar to, but not exactly the same as having a single large file system span all the data disks.

Hope this helps.

Okay I've looked through these forums, because I'm considering getting unRAID for an upcoming media server build.  I had a couple questions pretty similar to this poster's. 

I currently have one 500 gb drive which has the following directory setup:

F:\Pictures\

F:\Videos\

F:\Music\

I would like to keep this format (ie basically the same thing the original poster asked, to have it appear as if it was one drive).  However, the only method which I have seen require me to keep creating shortcuts for every single file I have, and then when I record new shows, I have to add more, which means this method is out.  I also seem to recall a method in which it would be possible to mount all of the hard drives to one folder.  It seemed from the explanation though that that would give me this type of file structure for example:

//tower/videos/disk1/movie1.ts

//tower/videos/disk2/movie2.ts

Is there any explanation someone could link me to that would give me:

//tower/videos/movie1.ts

//tower/videos/movie2.ts

If this is possible, how would writing to the drives be handled anyway, would the system just automatically write to the drive with the most space available, or would I have to set that up on my own with each recording?  My last question is a little bit off topic, but I saw something on the forums which said with unRAID, there is reduced write performance compared to RAID since it can't write on multiple disks.  Is it still the same write performance as a single disk, or is it reduced from that as well?  I'm trying to figure out if this or RAID5 is better for me.  I really like the fact that with unRAID, if multiple hard drives die, I only lose the data that was on those drives.  However, if I have to personally manage which recordings save where, RAID5 would probably be better for me.  My goal is to be able to just pick a directory that spans multiple drives and use Media Center to constantly write to that, until the whole span of drives runs out of space. 

This would probably be with two ATSC tuners and one analog tuner, so the write performance of RAID5 might be necessary?  I'm not really sure.  It would probably be fine as long as all of the tuners were writing to different drives anyway I guess.  Anyway, any information that could be given would make my decision that much easier.  Thanks.

I'm just getting started with UnRaid myself.  I was hoping to start this weekend.  I have 4 drives which includes the parity drive.  I would like to setup the drives so when viewing them from Windows I see ONLY 2 drives(shares), one for Music and Pics, and one for Movies. 

Disk 1 Music and Pictures

Disk2 and 3 Movies

I was hoping someone could point me to directions on how to setup the disks in this fashion. 

Also I hope someone can anwer this question.  CAn I setup UnRaid starting only with a Parity Drive and one Data drive?  I wanted to start this way before upgrading to the 6 drive package.

Also, I heard that it was better to create numerous 40GB partitions rather than using the whole drive as one partition.  Is this possible with UnRAid, should I follow this advice and how do I do this?

Thanks

Okay I've looked through these forums, because I'm considering getting unRAID for an upcoming media server build.  I had a couple questions pretty similar to this poster's. 

 

I currently have one 500 gb drive which has the following directory setup:

 

F:\Pictures\

F:\Videos\

F:\Music\

 

I would like to keep this format (ie basically the same thing the original poster asked, to have it appear as if it was one drive).  However, the only method which I have seen require me to keep creating shortcuts for every single file I have, and then when I record new shows, I have to add more, which means this method is out.  I also seem to recall a method in which it would be possible to mount all of the hard drives to one folder.  It seemed from the explanation though that that would give me this type of file structure for example:

 

//tower/videos/disk1/movie1.ts

//tower/videos/disk2/movie2.ts

 

Is there any explanation someone could link me to that would give me:

 

//tower/videos/movie1.ts

//tower/videos/movie2.ts

 

If this is possible, how would writing to the drives be handled anyway, would the system just automatically write to the drive with the most space available, or would I have to set that up on my own with each recording?

Yes this is (almost) possible using 'User Shares'.  "Almost" because in version 3.0 there is a "read-only" limitation which will be solved in upcoming version 3.1.  When we get User Shares completed in 3.1 there will be an accompying documention page.  But here's a brief description...

The 'default' method of exporting the data disks is to have each one exported as a separate share.  For example, if you have a system named 'tower' with 3 data disks, then you would have 3 shares:

//tower/disk1

//tower/disk2

//tower/disk3

When you navigate to one of the disk shares via My Network Places you would see all the files and directories in the 'root' of that disk.

When you enable User Shares, then the server will export another set of shares in addition to the normal disk shares.  The names of the additional shares are automatically generated simply by exporting each of the top-level directories on all the data disks, using the directory names for the share names.

For example, suppose you had the following directories on your 3 data disks:

disk1/Photos

disk1/Music

disk2/Video

disk3/Backup

When User Shares are enabled, unRAID Server will export these additional shares:

Photos

Music

Video

Backup

In this situation if you created a file in the Music share, then it will actually be created in disk1/Music directory.

Ok, so far so good, but what if the same top-level directory appears in more than one data disk?  In this case, the corresponding share will list the combined set of files/directories as if they were on one disk.  For example, suppose we have this:

disk1/Videos/episode1.ts

disk2/Vidoes/episode2.ts

disk3/Videos/episode3.ts

In this case the Videos share will produce this directory listing:

Videos/episode1.ts

Vidoes/episode2.ts

Videos/episode3.ts

When you go to read episode2.ts, then only disk2 will spin up during the transfer.

A limitation in version 3.0 is that you can not write the top-level Videos directory, however.  If you wanted to store episode4, then you would neeed to decide which disk to put it on, write to that disk via the disk share, and then click the "ReScan" button on the Shares page of the Management Utility.

This limitation is solved in upcoming version 3.1.

  My last question is a little bit off topic, but I saw something on the forums which said with unRAID, there is reduced write performance compared to RAID since it can't write on multiple disks.  Is it still the same write performance as a single disk, or is it reduced from that as well?  I'm trying to figure out if this or RAID5 is better for me.  I really like the fact that with unRAID, if multiple hard drives die, I only lose the data that was on those drives.

It's not that it can't write multiple disks - since there's a single parity disk, it becomes a bottleneck when writing multiple disk streams.  In addition, read/modify/write required for each write operation (to update parity) results in write-throughput in the range of 25%-40% speed of a single disk.  (Which BTW you would see in a highly fragmented RAID-5 system as well.)

  However, if I have to personally manage which recordings save where, RAID5 would probably be better for me.  My goal is to be able to just pick a directory that spans multiple drives and use Media Center to constantly write to that, until the whole span of drives runs out of space. 

Yes User Shares will accomplish this in version 3.1.

This would probably be with two ATSC tuners and one analog tuner, so the write performance of RAID5 might be necessary?  I'm not really sure.  It would probably be fine as long as all of the tuners were writing to different drives anyway I guess.  Anyway, any information that could be given would make my decision that much easier.  Thanks.

I think there are guys here on the forum who can successfully write multiple tuner streams, and read at the same time.  At present time, we don't know what the limitation is.

I'm just getting started with UnRaid myself.  I was hoping to start this weekend.  I have 4 drives which includes the parity drive.  I would like to setup the drives so when viewing them from Windows I see ONLY 2 drives(shares), one for Music and Pics, and one for Movies. 

Disk 1 Music and Pictures

Disk2 and 3 Movies

 

I was hoping someone could point me to directions on how to setup the disks in this fashion. 

This can be accomplished by setting up your disks like this:

disk1/Music

disk1/Pictures

disk2/Movies

disk3/Movies

Then enable User Shares on the Shares page in the Management Utility.  (and note current limitations described in previous post).

Also I hope someone can anwer this question.  CAn I setup UnRaid starting only with a Parity Drive and one Data drive?  I wanted to start this way before upgrading to the 6 drive package.

Sure you can have just 1 data drive.  You could also have a JBOD by simply not assigning a parity disk.

Also, I heard that it was better to create numerous 40GB partitions rather than using the whole drive as one partition.  Is this possible with UnRAid, should I follow this advice and how do I do this?

It's not possible to partition the unRAID data disks.  I'm not sure why you would want "numerous 40GB partitions", can you elaborate?

this sounds awesome!!  how can i be notified when 3.1 comes out?

also if i have user shares enabled and i have a movies share on 5 of my drives.  if i write a file to the movies share - how will unraid decide which of the 5 drives to write the file to?

Okay I've looked through these forums, because I'm considering getting unRAID for an upcoming media server build.  I had a couple questions pretty similar to this poster's. 

 

I currently have one 500 gb drive which has the following directory setup:

 

F:\Pictures\

F:\Videos\

F:\Music\

 

I would like to keep this format (ie basically the same thing the original poster asked, to have it appear as if it was one drive).  However, the only method which I have seen require me to keep creating shortcuts for every single file I have, and then when I record new shows, I have to add more, which means this method is out.  I also seem to recall a method in which it would be possible to mount all of the hard drives to one folder.  It seemed from the explanation though that that would give me this type of file structure for example:

 

//tower/videos/disk1/movie1.ts

//tower/videos/disk2/movie2.ts

 

Is there any explanation someone could link me to that would give me:

 

//tower/videos/movie1.ts

//tower/videos/movie2.ts

 

If this is possible, how would writing to the drives be handled anyway, would the system just automatically write to the drive with the most space available, or would I have to set that up on my own with each recording?

Yes this is (almost) possible using 'User Shares'.  "Almost" because in version 3.0 there is a "read-only" limitation which will be solved in upcoming version 3.1.  When we get User Shares completed in 3.1 there will be an accompying documention page.  But here's a brief description...

 

The 'default' method of exporting the data disks is to have each one exported as a separate share.  For example, if you have a system named 'tower' with 3 data disks, then you would have 3 shares:

 

//tower/disk1

//tower/disk2

//tower/disk3

 

When you navigate to one of the disk shares via My Network Places you would see all the files and directories in the 'root' of that disk.

 

When you enable User Shares, then the server will export another set of shares in addition to the normal disk shares.  The names of the additional shares are automatically generated simply by exporting each of the top-level directories on all the data disks, using the directory names for the share names.

 

For example, suppose you had the following directories on your 3 data disks:

 

disk1/Photos

disk1/Music

disk2/Video

disk3/Backup

 

When User Shares are enabled, unRAID Server will export these additional shares:

 

Photos

Music

Video

Backup

 

In this situation if you created a file in the Music share, then it will actually be created in disk1/Music directory.

 

Ok, so far so good, but what if the same top-level directory appears in more than one data disk?  In this case, the corresponding share will list the combined set of files/directories as if they were on one disk.  For example, suppose we have this:

 

disk1/Videos/episode1.ts

disk2/Vidoes/episode2.ts

disk3/Videos/episode3.ts

 

In this case the Videos share will produce this directory listing:

 

Videos/episode1.ts

Vidoes/episode2.ts

Videos/episode3.ts

 

When you go to read episode2.ts, then only disk2 will spin up during the transfer.

 

A limitation in version 3.0 is that you can not write the top-level Videos directory, however.  If you wanted to store episode4, then you would neeed to decide which disk to put it on, write to that disk via the disk share, and then click the "ReScan" button on the Shares page of the Management Utility.

 

This limitation is solved in upcoming version 3.1.

 

  My last question is a little bit off topic, but I saw something on the forums which said with unRAID, there is reduced write performance compared to RAID since it can't write on multiple disks.  Is it still the same write performance as a single disk, or is it reduced from that as well?  I'm trying to figure out if this or RAID5 is better for me.  I really like the fact that with unRAID, if multiple hard drives die, I only lose the data that was on those drives.

It's not that it can't write multiple disks - since there's a single parity disk, it becomes a bottleneck when writing multiple disk streams.  In addition, read/modify/write required for each write operation (to update parity) results in write-throughput in the range of 25%-40% speed of a single disk.  (Which BTW you would see in a highly fragmented RAID-5 system as well.)

 

  However, if I have to personally manage which recordings save where, RAID5 would probably be better for me.  My goal is to be able to just pick a directory that spans multiple drives and use Media Center to constantly write to that, until the whole span of drives runs out of space. 

Yes User Shares will accomplish this in version 3.1.

 

This would probably be with two ATSC tuners and one analog tuner, so the write performance of RAID5 might be necessary?  I'm not really sure.  It would probably be fine as long as all of the tuners were writing to different drives anyway I guess.  Anyway, any information that could be given would make my decision that much easier.  Thanks.

I think there are guys here on the forum who can successfully write multiple tuner streams, and read at the same time.  At present time, we don't know what the limitation is.

this sounds awesome!!  how can i be notified when 3.1 comes out?

Easiest way is to sign up for notification of new posts to the Announcements topic.

also if i have user shares enabled and i have a movies share on 5 of my drives.  if i write a file to the movies share - how will unraid decide which of the 5 drives to write the file to?

Current algorithm is that it writes across them more-or-less "evenly".  In a nutshell: it will pick a disk and keep writing to it until free space is below some threshold, at which point it will pick another disk which still has an amount of space still above threshold and start writing to it.  When no more disks remain with space above threshold, the threshold is lowered.  There are more details which will be described when we release 3.1 "final".

sounds good - thanks!