I'm setting up a test of unRaid this weekend to try it out.  I emailed support a couple of times hoping to have a couple of questions answered, but have gotten no response so far.  Hopefully someone here might be able to either answer or point me to some answers.

 

1. I am currently running WHS2011 with 11 drives (that includes the boot drive) - so effectively, only 10 are used for storage.  Since unRaid boots on a flash drive, that means that (if I purchase unRaid) I should be able to use all 11 drives for storage - is that correct?

 

2. If I want a parity drive, the size of the parity drive needs to be equal or greater than any other drive I am using - correct?

 

3. I currently use a variety of USB drives as backup devices.  Will I still be able to do that with unRaid?  I found some references to S.N.A.P. in the forums, but haven't found anything that clears up that question for me.

 

I have several more questions, but I'm hoping that doing a basic setup and trying things will help me find the answers for the rest of my questions.

 

Thanks in advance for any help.

 

~Whip

1. Correct but you need to think about how to transfer from one to the other.  You can't just add te drives to unRAID with the data intact.  Each added drive will be wiped out.  You have to copy over.

 

2. Correct.  Parity has to be at least as large as any data drive.

 

3.  You can't use USB for a parity protected drive.  You can have a drive connected via USB using Snap to make it visible, but it will not be parity protected.  I am not sure if it will work with user shares or not.

1 is NOT correct! While he may not need a boot drive with unRAID he will need a parity drive to have any protection and that will eat a drive. With unRAID he will still have only 10 storage drives. Since parity is restricted to just one drive he may still gain overall storage due to reduced overhead anyway.

 

Transferring the data over is indeed a problem if he intends to use the existing hardware. Chicken and egg I'm afraid....

You can use SNAP and usb drives to backup data stored in unRAID to the USB drives. The USB drives can be exported by unRAID as individual shares but they cannot be part of the protected array.

1 is NOT correct! While he may not need a boot drive with unRAID he will need a parity drive to have any protection and that will eat a drive. With unRAID he will still have only 10 storage drives. Since parity is restricted to just one drive he may still gain overall storage due to reduced overhead anyway.

 

Transferring the data over is indeed a problem if he intends to use the existing hardware. Chicken and egg I'm afraid....

 

Yes you are correct.  I guess I was considering parity as part of the "storage"... But technically it's not.  I think he understood that one drive is for parity only based on his questions.

 

Did not intend to be misleading.

 

You need at least one two spare drives for the copy, over and above your total data.  If you copy over one drive at a time and keep increasing the array, it will be very slow.  I think it took me 3 rounds of copy / increase array to get mine complete.

Thanks a ton.  That's what I was looking for.

 

As for the USB's, I use them as non-live backups.  Separate from my server.  They are intended to be my recovery if I lose my server.  I would like to continue to use them as such (I think rsync is something that I could use to do scheduled backups).

 

I was hoping that I would be able to attach each of my USB's and copy from them to storage in a running unRaid server.

 

Thanks again.

 

~Whip

I was hoping that I would be able to attach each of my USB's and copy from them to storage in a running unRaid server.

 

Yes you could do that.

 

Thanks for all the information, its been a great help.

 

So I've setup my free unRaid server.  Here are the specs:

 

MSI 880GM-E43 AM3 AMD 880G HDMI Micro ATX AMD Motherboard

AMD Sempron 140 Sargas 2.7GHz Socket AM3 45W Single-Core Processor SDX140HBGQBOX

4 x Mushkin Enhanced Essentials 2GB 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10666) Desktop Memory Model 991586

4 x SAMSUNG EcoGreen F4 ST2000DL004 2TB SATA 3.0Gb/s 3.5" Internal Hard Drive -Bare Drive

1x Hitachi HDT721010SLA360 1TB SATA

 

The issue I ran into is that only 3 of the Samsungs and the Hitachi showed up as available under devices.  Because I wanted to get something setup to help me learn about unRaid, I went ahead and selected the Samsungs as my drives (parity and 2 data drives).  I started the preclear_disk running against the Hitachi to see how long that takes (still going).

 

Which leads me to my first question: 

1.Is there a way to determine which disk is not being seen by unRaid (other than removing them 1 at a time and testing).

 

Also leads to a few more thoughts: 

 

Hardware compatibility:

 

I found a thread here on my MB so it looks like it should be fine - unless I've missed something else.

 

I also have a Rosewill esata external enclosure with 5x 1.5 TB drives in it that I want to add to system.  I can use the onboard esata port or the rosewill 2 port esata card that came with the system.

 

2. Since my MB appears to be Level 1 certified, I'm assuming that means the esata on it should be good to use (even though I've found nothing specific on using the esata port on my MB) - Am I thinking about this correctly?

 

Shares:

 

Based on what I've read it looks like I should enable user shares and create my shares now.  However, I noticed a couple of threads on "Split level" issues.  I will have shares for "Movies" "TV" "Photos" "Music" & "Videos". 

 

3. If I create them now, before I've added all of my drives, will I run into an issue if I set them with a Split Level of 1 as I continue to add my drives?

 

Preclear:

 

So last area of questions - reading the configuration guide(s) that I've run across, I'm under the impression that using the preclear script will keep me from taking down my entire array when I add a new drive.  I understand that if I don't use the preclear script, then unRaid will do effectively the same thing when I add a new drive but unRaid takes down the entire array while it performs this action.

 

4. Is my understanding of this correct?

 

If it is, then I'm confused.  Why didn't unRaid do that when I added my drives this weekend when I first setup my system (I'm using the free version atm)?

 

Again, thanks for all the help.  I'm actively searching for guides, FAQs etc, so if I've missed something out there, please don't hesitate to point it out.

 

I appreciate everything.

 

~Whip

Thanks for all the information, its been a great help.

 

So I've setup my free unRaid server.  Here are the specs:

 

MSI 880GM-E43 AM3 AMD 880G HDMI Micro ATX AMD Motherboard

AMD Sempron 140 Sargas 2.7GHz Socket AM3 45W Single-Core Processor SDX140HBGQBOX

4 x Mushkin Enhanced Essentials 2GB 240-Pin DDR3 SDRAM DDR3 1333 (PC3 10666) Desktop Memory Model 991586

4 x SAMSUNG EcoGreen F4 ST2000DL004 2TB SATA 3.0Gb/s 3.5" Internal Hard Drive -Bare Drive

1x Hitachi HDT721010SLA360 1TB SATA

 

The issue I ran into is that only 3 of the Samsungs and the Hitachi showed up as available under devices.  Because I wanted to get something setup to help me learn about unRaid, I went ahead and selected the Samsungs as my drives (parity and 2 data drives).  I started the preclear_disk running against the Hitachi to see how long that takes (still going).

 

Which leads me to my first question: 

1.Is there a way to determine which disk is not being seen by unRaid (other than removing them 1 at a time and testing).

Stop the array, which disk is NOT listed in the drop-down lists on the device assignment page?  (they are listed by model/serial number)  The one not listed is the one not recognized/plugged in correctly.  (You should note the serial number of each drive as you install them.  Many write the serial number on a sticker on the drive itself if not visible when mounted.  You can also type on the command line:

ls -l /dev/disk/by-id

the disks and their partitions will be listed there, again by their model/serial numbers.

 

Also leads to a few more thoughts: 

 

Hardware compatibility:

 

I found a thread here on my MB so it looks like it should be fine - unless I've missed something else.

 

I also have a Rosewill esata external enclosure with 5x 1.5 TB drives in it that I want to add to system.  I can use the onboard esata port or the rosewill 2 port esata card that came with the system.

 

2. Since my MB appears to be Level 1 certified, I'm assuming that means the esata on it should be good to use (even though I've found nothing specific on using the esata port on my MB) - Am I thinking about this correctly?

Can't help you there. I have entirely different hardware

Shares:

 

Based on what I've read it looks like I should enable user shares and create my shares now.  However, I noticed a couple of threads on "Split level" issues.  I will have shares for "Movies" "TV" "Photos" "Music" & "Videos". 

 

3. If I create them now, before I've added all of my drives, will I run into an issue if I set them with a Split Level of 1 as I continue to add my drives?

No, the split level can be changed at any time.

 

Preclear:

 

So last area of questions - reading the configuration guide(s) that I've run across, I'm under the impression that using the preclear script will keep me from taking down my entire array when I add a new drive.  I understand that if I don't use the preclear script, then unRaid will do effectively the same thing when I add a new drive but unRaid takes down the entire array while it performs this action.

 

4. Is my understanding of this correct?

Yes

If it is, then I'm confused.  Why didn't unRaid do that when I added my drives this weekend when I first setup my system (I'm using the free version atm)?

The "clearing" of a drive by unRAID to maintain parity is ONLY done after you add a parity drive to the array and parity is initially established.    Since parity was not already calculated, there was no need to ensure the disks were zeroed.

 

The other major purpose of the preclear script is to get the disks past the early part of the "bathtub curve"  (google it) in failures.  Roughly 1 in 5 drives seems to arrive new from the manufacturer with problems.  Do you feel lucky?  Many do not want to risk their data.

Stop the array, which disk is NOT listed in the drop-down lists on the device assignment page?  (they are listed by model/serial number)  The one not listed is the one not recognized/plugged in correctly.  (You should note the serial number of each drive as you install them.  Many write the serial number on a sticker on the drive itself if not visible when mounted.  You can also type on the command line:

ls -l /dev/disk/by-id

the disks and their partitions will be listed there, again by their model/serial numbers.

 

Awesome, thanks.

 

No, the split level can be changed at any time.

 

Good, that will make things simpler.

 

The "clearing" of a drive by unRAID to maintain parity is ONLY done after you add a parity drive to the array and parity is initially established.    Since parity was not already calculated, there was no need to ensure the disks were zeroed.

 

The other major purpose of the preclear script is to get the disks past the early part of the "bathtub curve"  (google it) in failures.  Roughly 1 in 5 drives seems to arrive new from the manufacturer with problems.  Do you feel lucky?  Many do not want to risk their data.

 

Clarity!  Thanks a ton.

 

I'm not using any new hardware.  I'm converting my WHS2011 server that has been in service for almost 1.5 years.  Because of that I have very little fear of the drives failing prematurely.  Because of this, would it possibly make sense to add all my drives before adding my parity drive (seeing as they are all going to be empty)?  Then as I add new drives I would preclear them before adding to the array.  (Frankly, the idea of preclearing all 7 remaining drives is really depressing)

 

Thanks for the continued help!

 

~Whip

I'm not using any new hardware.  I'm converting my WHS2011 server that has been in service for almost 1.5 years.  Because of that I have very little fear of the drives failing prematurely.  Because of this, would it possibly make sense to add all my drives before adding my parity drive (seeing as they are all going to be empty)?  Then as I add new drives I would preclear them before adding to the array.  (Frankly, the idea of preclearing all 7 remaining drives is really depressing)

 

Thanks for the continued help!

 

~Whip

I would get a smart report on each of your existing drives.  If there are zero sectors re-allocated, and zero sectors pending re-allocation, I would not worry as much either.  (you are past the initial part of the bathtub curve)  If there are sectors pending re-allocation, or already re-allocated, I'd run a single pass of the preclear script... just to be sure the issues are not ongoing.

 

Joe L.

I would get a smart report on each of your existing drives.  If there are zero sectors re-allocated, and zero sectors pending re-allocation, I would not worry as much either.  (you are past the initial part of the bathtub curve)  If there are sectors pending re-allocation, or already re-allocated, I'd run a single pass of the preclear script... just to be sure the issues are not ongoing.

 

Joe L.

 

I'll do that, thanks again!

 

~Whip

I also have a Rosewill esata external enclosure with 5x 1.5 TB drives in it that I want to add to system.  I can use the onboard esata port or the rosewill 2 port esata card that came with the system.

 

This will be extremely slow. eSATA can only support 2 drives per port with adequate performance. ThermalTake makes inexpensive 2 bay eSATA docks.

 

 

This will be extremely slow. eSATA can only support 2 drives per port with adequate performance. ThermalTake makes inexpensive 2 bay eSATA docks.

 

The Rosewill enclosure uses port multipliers? I think.  I will say that there has never seemed to be a performance issue when copying to the enclosure vs. copying to the internal drives.  Under Win 7 copying to the server running WHS2011 I would get consistent copy speeds of around 45mb(? I'm not sure of MB/S or mbs or what) regardless of which drives I was hitting.  I don't know if that is slow or not, just I know that it's been consistent.

 

All that said, if changing enclosures would make it faster, I'll look into it.

 

One more question - SmartCtl (everything I've found indicates this will give me a smart report).  I've looked up the manpage for it and I'm reminded of why Linux and I don't get along.  I feel like I'm not reading English.  Can someone give me an example of how to get a smart report from smartctl?  I know that one of my drives is /dev/sbd, for example.

 

~Whip

To run a smart test on sdb you would run:

smartctl --test=short /dev/sdb

 

It takes ~ 1 minute to complete, after that, to view results:

smartctl -a /dev/sdb

 

 

To run a smart test on sdb you would run:

smartctl --test=short /dev/sdb

 

It takes ~ 1 minute to complete, after that, to view results:

smartctl -a /dev/sdb

 

Thanks!

 

~Whip

First, I just wanted to thank everyone for the help.

 

Second, I spent the remainder of the weekend moving drives around, getting serial numbers off of them, running smart tests, etc. 

 

As a last note, unRaid seems to recognize the port multiplier on my rosewill enclosure.  In the end though, I do want to put as many of the drives in my actual tower as possible.  I will need to add a controller card to do that.  Since I will need 5 more ports internally to "fill" my tower, I think the SuperMicro AOC-SASLP-MV8 seems to be popular solution around here.

 

I am re-reading many posts/faqs because things make a great deal more sense to me now that I've messed around a bit with unRaid.  I think I'm sold.  Well, actually I know I'm sold.  I will be picking up my pro license(s) this week.

 

~Whip

 

 

This will be extremely slow. eSATA can only support 2 drives per port with adequate performance. ThermalTake makes inexpensive 2 bay eSATA docks.

 

The Rosewill enclosure uses port multipliers? I think.  I will say that there has never seemed to be a performance issue when copying to the enclosure vs. copying to the internal drives.  Under Win 7 copying to the server running WHS2011 I would get consistent copy speeds of around 45mb(? I'm not sure of MB/S or mbs or what) regardless of which drives I was hitting.  I don't know if that is slow or not, just I know that it's been consistent.

 

All that said, if changing enclosures would make it faster, I'll look into it.

 

One more question - SmartCtl (everything I've found indicates this will give me a smart report).  I've looked up the manpage for it and I'm reminded of why Linux and I don't get along.  I feel like I'm not reading English.  Can someone give me an example of how to get a smart report from smartctl?  I know that one of my drives is /dev/sbd, for example.

 

~Whip

 

unRAID need concurrent access to all array drives. More than 2 drives in the enclosure will make parity checks and disk rebuilds take several days.

unRAID need concurrent access to all array drives. More than 2 drives in the enclosure will make parity checks and disk rebuilds take several days.

 

Good to know the details - and I've only got 2 drives in the enclosure now