RAID Options

[aadeshgandhi]

Hello,
I am new to RAID and would like to get your opinion on what my options are.
Here is what I have
1) Old Dell PowerEdge SC1420 with CentOS 6.4 installed on it.
2) Presently running with a 36GB SATA HDD.
3) Has Plex installed on it.
4) About 1.5TB of media that is either located on windows laptop or external HDD's

Here is what I want to do.
1) Move all this data on the desktop
I have ordered a 3TB internal SATA drive and had thought that once that fills up I can order another one. Problem is these will show up as two drives instead of one 6TB drive. That's where RAID comes in.

I am new to this and not sure what my options are.. should i go for a RAID Controller or a RAID enclosure?? Which is better and why
What are recomendations for both? I would like to have 5 ports. Budget is $200.


Thanks,
Aadesh.

 

[dave_the_nerd]

For a home streaming server, software RAID is fine - you shouldn't need an extra controller.

You'll want to set up the RAID before you start copying data to it. (Doing it afterwards is... detrimental to the data already on the drive.) So if you want to go the RAID route, you'll ned to buy your drives up front.

How are you backing up?

Given your budget, you don't have a lot of choices. If you want to add more storage in later (more economical anyway), and if you have some kind of backup system working, and if you have to have it present your media to the plex server as a "single" drive, you're probably better off just using symlinks or mounting the new drive to a folder in the old drive's folder structure.

 

[aadeshgandhi]

How are you backing up?

Right now I am not.. just have the data scattered on the laptop and external HDD

For a home streaming server, software RAID is fine - you shouldn't need an extra controller.

Would the software RAID be stable?

Given your budget, you don't have a lot of choices. If you want to add more storage in later (more economical anyway), and if you have some kind of backup system working, and if you have to have it present your media to the plex server as a "single" drive, you're probably better off just using symlinks or mounting the new drive to a folder in the old drive's folder structure.

Yes, I would like to add more drives later, right now I have two 3TB that I can get started with because a friend told me that I would need at least two drives to do RAID and then I can add extra drives later. Essentially I am looking at as you said one single drive of 10TB atleast hence I said number of ports to be 5. Could you explain me more on how to use symlinks and mounting the new drive to a folder in the old drive's folder structure.


Thanks,
Aadesh

 

[thecoolnessrune]

I know this gets said alot but I think this needs said again because it seems you are going down this path.

RAID is not a backup, it is for high availability.

If you are getting into RAID in order to have a backup, then you are not going about this correctly and should instead be looking into a separate backup solution.

If you are looking into RAID for high availability (not having to restore from backup when a hard drive dies) then you need to address your concerns:

How important is speed?
How important is it that the server does not go down?
How much storage do you need vs how much you are willing to buy?

And if you want an expandable RAID you can't use any traditional RAID. Traditional RAID (RAID 1, 5, 10, etc.) are old storage technologies that do not allow any sort of expansion of Storage volumes.

If you want a flexible Soft RAID, then consider using something like unRAID which, while not high performance, is suitable for home use. It allows you to dynamically add and remove drives from the pool and you can sustain 1 drive failure without the system going down.

Again, please keep in mind that RAID is not a backup.

 

[XavierMace]

I know this gets said alot but I think this needs said again because it seems you are going down this path.
And if you want an expandable RAID you can't use any traditional RAID. Traditional RAID (RAID 1, 5, 10, etc.) are old storage technologies that do not allow any sort of expansion of Storage volumes.

That depends on the controller. Most actual RAID controllers support online array expansion.

 

[smitbret]

edit.... got too convoluted

 

[thecoolnessrune]

That depends on the controller. Most actual RAID controllers support online array expansion.

Expansion, but none that I know of provide for Shrinking of an array, and even fewer that support online expansion properly re-balance the array once its been grown, leading to a false understanding of the redundancy of the array.

 

[smitbret]

Hello,
I am new to RAID and would like to get your opinion on what my options are.
Here is what I have
1) Old Dell PowerEdge SC1420 with CentOS 6.4 installed on it.
2) Presently running with a 36GB SATA HDD.
3) Has Plex installed on it.
4) About 1.5TB of media that is either located on windows laptop or external HDD's

Here is what I want to do.
1) Move all this data on the desktop
I have ordered a 3TB internal SATA drive and had thought that once that fills up I can order another one. Problem is these will show up as two drives instead of one 6TB drive. That's where RAID comes in.

I am new to this and not sure what my options are.. should i go for a RAID Controller or a RAID enclosure?? Which is better and why
What are recomendations for both? I would like to have 5 ports. Budget is $200.


Thanks,
Aadesh.

Let me see if I understand.

You have your PowerEdge lying around with a 36GB drive in it and would like to add a 3TB HDD to it and use it just as a server so you can consolidate files from a bunch of external HDDs and a laptop. Correct?


Then, when that drive is full, you would like to add a 2nd 3TB HDD and have them presented as one 6TB HDD. Also correct? This is not RAID. This is called drive pooling. You could set up both of those drives under RAID 0 and it would show up this way but you would need to start with 2 empty drives, regardless of whether it was hardware or software RAID. You could also get an add-in card that supports JBOD (Just a Bunch of Drives). All other RAID or parity options will require a loss of storage space and/or more HDDs.

How big are these external HDDs and would you be able to use them in the server?

 

[aadeshgandhi]

Let me see if I understand.

You have your PowerEdge lying around with a 36GB drive in it and would like to add a 3TB HDD to it and use it just as a server so you can consolidate files from a bunch of external HDDs and a laptop. Correct?


Then, when that drive is full, you would like to add a 2nd 3TB HDD and have them presented as one 6TB HDD. Also correct? This is not RAID. This is called drive pooling. You could set up both of those drives under RAID 0 and it would show up this way but you would need to start with 2 empty drives, regardless of whether it was hardware or software RAID. You could also get an add-in card that supports JBOD (Just a Bunch of Drives). All other RAID or parity options will require a loss of storage space and/or more HDDs.

How big are these external HDDs and would you be able to use them in the server?

Smitbret,
I mentioned RAID cause although the data is not critical I have spent time gathering it so I just wanted to have a way to save the data if one drive fails. I understand that this is not back-up.... my whole point over here is to have a folder which has 3TB to start with and as I add new drives all that space is also available for the said folder.. right now i have two 3TB drives which are empty...
How is JBOD different than RAID and would it be a better option than RAID?

 

[thecoolnessrune]

Smitbret,
I mentioned RAID cause although the data is not critical I have spent time gathering it so I just wanted to have a way to save the data if one drive fails. I understand that this is not back-up.... my whole point over here is to have a folder which has 3TB to start with and as I add new drives all that space is also available for the said folder.. right now i have two 3TB drives which are empty...
How is JBOD different than RAID and would it be a better option than RAID?

RAID = Redundant Array of Independent Disks
JBOD = Just a Bunch of Disks

You can kind of see the difference in the name A RAID array is meant to provide redundancy of data in case of a drive failure, which is not a backup, but rather just high availability if one drive goes down. This is why RAID does not "save the data if one drive fails"

• RAID doesn't protect you against a file being deleted.
• RAID doesn't protect you against a file being overwritten.
• RAID doesn't protect you from your system being compromised and all of your data being overwritten, deleted, or corrupted.
• RAID doesn't protect you from URE's (Unrecoverable Errors that you are unlikely to find until you are trying to rebuild an array)
• RAID doesn't protect you from Catastrophic damage (someone dumps water onto the server)
• RAID doesn't protect you from Virus' or Software bugs that wipe out data

I could go on, but I hope that is enough to make it understandable. In fact Journalspace.com and Magnolia.com are great examples of what happens when people confuse RAID as a backup.

Keep in mind that its like a ladder. RAID <- Backup <- Archive. An archive is a backup, and a backup is redundant, but a RAID is not a backup, and a backup is not automatically an archive.

To continue answering your question, a JBOD is just a bunch of disks. All it does is bring a bunch of disks together into one giant, accessible storage pool without any sort of redundancy or parity. Drive pooling creates dynamic JBODs that leverage the power of traditional JBODs (dynamically adding and removing storage, and making the pool bigger without having to manage an array), while also providing a varying number of features.

Windows Home Server version 1 used something called "Tombstones" to make data in a drive pool semi-redundant. It did this by writing all incoming data to a "Landing Pad" and then distributing it in as many copies as required by the user onto the drive array, making sure the data was located on a different drive. Other softwares like unRAID, FlexRAID, and SnapRAID use similar systems (though without landing pads now) to create the appearance of redundancy in a drive pool, and also lets the data be accessible on individual drives in case there is a massive hardware failure that takes out several hard drives or other computer components.

 

[aadeshgandhi]

This is why RAID does not "save the data if one drive fails"

wohaa... all of that was too technical for me... Here is my understanding of a RAID 5 conifg...

If I create a RAID 5 with 4 X 3TB disks, I essentially will be getting 9TB of useful space. If one drive fails, I can replace it and still my 9TB would be safe...

Please let me know if I am horribly wrong on this one.


Aadesh.

 

[smitbret]

Smitbret,
I mentioned RAID cause although the data is not critical I have spent time gathering it so I just wanted to have a way to save the data if one drive fails. I understand that this is not back-up.... my whole point over here is to have a folder which has 3TB to start with and as I add new drives all that space is also available for the said folder.. right now i have two 3TB drives which are empty...
How is JBOD different than RAID and would it be a better option than RAID?

JBOD of drives is like you have 4 pieces of different sized rope:

1 foot
4 feet
2 feet
3 feet

If you tie/splice them all together then you have 10 feet of rope. That's JBOD.

RAID is different and there are different types of RAID. I'll give you the most common examples, but Googling RAID or hitting it up in Wikipedia will be more thorough and probably more accurate.

RAID 0 is where you take 2 or more drives of equal size and spread the files across all of the drives. For example, if you have 3 drives in a RAID 0, the files would be broken up into 3 parts and each drive would have a part of it. Makes accessing data much faster because you can access from all 3 drives at the same time. It's also very risky because if one drive dies, you lose ALL of your data because 1/3 of every file is missing. The more drives you have the more risky it gets because 1 dead drive kills everything. You have to start with empty drives because everything gets erased on all the drives when you set it up. Adrenaline junkies like this type but it is not safe at all.

RAID 1 is where you have 2 (or 4 or 6, etc.) drives set up to mirror each other. You have two 3TB drives but they are mirrors of each other. When you save a file, it saves an exact copy to each drive. If one drive fails, the other takes over. It is quite safe, but won't protect you from things like accidental deletions or if a power surge fries everything in your PC.

RAID 5 is where you have 3 or more drives set up and the data is distributed in a way that allows any 1 drive to fail but you'll still be able to run and recover as long as you don't lose another and you lose the space of one of the drives. For example, you have five 1TB drives in a RAID 5. The space available would actually be 4TB instead of 5TB because of redundancy. It's slower to write because of space calculations but is safer than JBOD or RAID 0 but not as safe as RAID 1.

RAID 6 is like RAID 5 but with 2 drives setup for redundancy and so you could lose any 2 drives and be OK. Five 1TB drives would then give you 3TB of available space.

There are other types like 0+1 and RAID 10 but they aren't common for home use.

These RAID setups can be done either hardware or software, but the end results are usually the same. The difference is that true Hardware RAID doesn't steal CPU cycles to calculate Parity and Redundancy so it is faster than software where your PC's CPU takes care of it.

Then there are the newer versions that aren't technically RAID but follow the same ideas. ZFS is a new file system like FAT32 and NTSC. With ZFS you can set up RAIDz1 or RAIDz2 that are just like RAID 5 and RAID 6 except ZFS brings a lot of tools to the table to protect your data from corruption and bitrot. There are lots of iterations of this like FreeNAS and Solaris and most are free but they run on their own Linux distributions and can be a little limiting if you want to do anything besides just store and serve files. You also have to start with empty drives for this, too.

Then, there are the non-RAID, Parity options like unRAID, FlexRAID and SnapRAID that have their own "versions" of RAID 5 or 6. They are software and generally less expensive to implement than true RAID and ZFS solutions. They are also slower but if you don't need it to handle huge amounts of I/O requests and just want a centralized place to store photos, files, music and to serve up the occasional video to a normal home then they are plenty fast. The advantage of these is that you can continually add drives to them and expand your storage space.

With your limited budget, I would steer you in this direction, particularly towards unRAID.

Take that $200, buy an unRAID license, a 1GB flash drive and a 2nd 3TB HDD. One of your 3TB HDDs will be your parity drive. Every drive after that can be used as a storage space and you can continually add on for as many drives as you can stuff in your PowerEdge and it doesn't matter how big they are. The OS actually runs off the flash drive. No HDD needed.

From the info you've told us, you would then have two 3Tb and a 36GB drive. In an unRAID setup you would then have ~3036GB. Any drives you add to it would just be additional space. You are just limited by the size of your parity drive that has to be the same size as your largest drive. If you wanted to add those externals to the array, just copy their data to the array you've already created, then clear 'em and add 'em in. If you had a couple of 500GB external HDDs then you'd have added another TB of storage and as soon as you rebuild the array it's all parity protected.

So if you wanted to start using 4TB drives then you would slip a new 4TB in as the parity drive and move the 3TB in to hold data. Your storage space would then be ~6036GB and if you added another 4TB you'd be at 10+ TB. Just keep adding drives as you need and your budget allows. You don't have to lose data or erase anything. You just have to clear the drives before you add them to the array.

unRAID also has a Plex plugin so you can stream your media, too. The other nice thing about unRAID is you can pretty much just use any old hardware that you have lying around. It is VERY light on resources, doesn't require much RAM and you would only need to consider CPU speed if Plex is transcoding your video files.

FlexRAID is similar but runs on top of an existing OS like Windows and allows you to import disks that have data on them. You'd also have to install the OS to a HDD. It also has its own version of RAID 6 if 1 parity disk doesn't make you feel safe enough. SnapRAID is very similar to FlexRAID and is open source but uses CLI for administration. There's a GUI for it called Elucidate but I never could get it to work quite right. I'm sure I'm just an idiot.

Sorry for the length, but you don't just either RAID or not. Lots of stuff to look into.

 

[smitbret]

wohaa... all of that was too technical for me... Here is my understanding of a RAID 5 conifg...

If I create a RAID 5 with 4 X 3TB disks, I essentially will be getting 9TB of useful space. If one drive fails, I can replace it and still my 9TB would be safe...

Please let me know if I am horribly wrong on this one.


Aadesh.

Correct. But to add another voice, backup anything you can't afford to lose even if it's in RAID.

I run FlexRAID for some security, but I backup all of my documents, music, photos, etc. The only thing I don't backup is my DVD and BR rips because they are so large and I already have the backups in the form of the DVDs and BRs.

 

[thecoolnessrune]

wohaa... all of that was too technical for me... Here is my understanding of a RAID 5 conifg...

If I create a RAID 5 with 4 X 3TB disks, I essentially will be getting 9TB of useful space. If one drive fails, I can replace it and still my 9TB would be safe...

Please let me know if I am horribly wrong on this one.


Aadesh.

This is how it works in theory. In reality a failed hard drive can be a varied experience. The biggest enemy of parity rebuilds (which is every RAID that does not have 1 in its name such as RAID 5 and 6), is URE's, or the Unrecoverable Error. This usually happens when data is silently corrupted on one or more hard drives in your array. In today's world of 3 and 4TB hard drives, this is actually very likely on consumer drives (like you're wanting to use). A single URE can screw your entire RAID pool because RAID runs on hard drive sectors not data. That means if a single sector is bad in a RAID 5 array and you've already lost a drive, you're screwed. This is made less likely by scrubbing your array, but this is not often done at the consumer level and difficult and time consuming to do with Soft Raid (Hardware RAID cards have the proper ECC memory and ASICs that make for proper scrubbing)

End result, RAID 5 does not belong on high capacity Consumer arrays, and you aren't doing RAID 6 in Soft Raid (I don't know of any chipsets that will do RAID 6 since the parity calculations are so complicated).

This is why a Soft Raid based on a Drive Pool, as already stated, is best for you. unRAID will let you have a form of a RAID 3 array, which is file-based storage with a dedicated parity drive. This sort of Soft Raid doesn't have to worry about URE's preventing an entire restore, but can still encounter corruption taking out individual files or groups of files. Again, backing up important files is key.

 

[aadeshgandhi]

Correct. But to add another voice, backup anything you can't afford to lose even if it's in RAID.

I run FlexRAID for some security, but I backup all of my documents, music, photos, etc. The only thing I don't backup is my DVD and BR rips because they are so large and I already have the backups in the form of the DVDs and BRs.

Yes, I will ensure that I have back-up of my important stuff... I want to create this massive folder so that I can make it my PLEX server also use that folder for my FTP site so that everything is listed there...

Having said that.. what do you feel.. should I go JBOD route or RAID5?

Also lets say I have 4 X 3TB in JBOD, what I got from discussion is it will create a 12TB folder, what will happen if a drive fails? I will loose all the data from that drive forever?

Also is drive failing fairly common? I have working with computers for many years now but never had once I seen my HDD fail.. OS... yes many a times but not the drive itself.. external one's too have not failed yet.. fingers crossed!

 

[smitbret]

That's another reason to go unRAID or FlexRAID. If you have a 2nd drive go down, you only lose the data on those drives. All other drive data is recoverable.

Traditional RAID, outside of RAID 1, is really just inferior to all the other options out now like ZFS, unRAID, etc.

 

[smitbret]

IMHO, I mean.

 

[thecoolnessrune]

Yes, I will ensure that I have back-up of my important stuff... I want to create this massive folder so that I can make it my PLEX server also use that folder for my FTP site so that everything is listed there...

Having said that.. what do you feel.. should I go JBOD route or RAID5?

Also lets say I have 4 X 3TB in JBOD, what I got from discussion is it will create a 12TB folder, what will happen if a drive fails? I will loose all the data from that drive forever?

Also is drive failing fairly common? I have working with computers for many years now but never had once I seen my HDD fail.. OS... yes many a times but not the drive itself.. external one's too have not failed yet.. fingers crossed!

Whether or not you lose all data or not on drive failure depends on how the JBOD is constructed as there is no set method.

Windows Home Server version 1 was notorious about its JBOD because it would store its tombstones on the host hard drive. If that hard drive failed it was extremely time consuming and difficult to replace the host hard drive and operating system without overwriting the tombstone links on the JBOD drives, thus destroying the data on the array.

Mac OSX's "Concatenated RAID" also has similar issues, in that since the Table of Contents of the JBOD was stored on the first disk, it was a bit of a russian roulette. If the first drive failed, you lost everything, but if any but the first drive in the JBOD set failed, you only lost the data on that drive.

The soft raids like unRAID, FlexRAID, and SnapRAID will only lose data on the lost hard drive. If you use parity (sacrifing one of your hard drives to be used as a parity drive), then you can lose any one drive and still keep all of your data.

So in your use case, if you buy 4 3TB drives and make a 12TB folder (which means you're not using parity), and use a Soft Raid like unRAID, you'll simply keep losing the data on every hard drive that fails. If you make a 9TB array and dedicate one of the 3TB drives as a parity drive, then you'll be able to lose one hard drive in the array without losing any data.

As to hard drives failing, there's all sorts of metrics on this but ultimately, it seems if it doesn't fail immediately, you tend to be fine (for the useful life of your server). There is all sorts of metrics that go into it though and home environments are worse than server environments regardless of what people think. Things that affect a hard drive:

Workload
Vibration
Spinup/Spindowns
Temperature variance

The workload in a hard drive is worse in a server environment, but everything else is worse in a home environment. Vibrations are worse. One of the big killers of hard drives is constant spin ups and spin downs, but since people want their server to be "cool and quiet", they have their drives constantly doing this, which kills them sooner. This is why server hard drives don't spin down (that and performance reasons). Lastly, a datacenter is kept at a constant, warm temperature with plenty of airflow. A hard drive shouldn't go over 50 degrees Celsius and its temperature should be kept *even*. This doesn't happen in home environments with often inadequate airflow, dust, and constant spin ups and spin downs that cycle the hard drive temperatures.

Anecdotally, I bought 6 3TB hard drives for my array (a zpool of 3 sets of mirrored disks). Within 48 hours, one of them failed and had to be replaced.

So take reliability as you will, but its certifiably silly to plan anything important in the technology realm around the idea that it probably won't fail in its lifetime. It can, and it will.

 

[XavierMace]

I'll stick with my hardware RAID. I've seen far too many software RAID setups fail and lose all the data to trust it with anything important. Plus, I don't like waiting for files to copy. You can get older model HP Smart Array's on Amazon for $100. A Smart Array P800/512mb supports more drives than you can shake a stick at and will outperform any software RAID setup that I've used.

 

[smitbret]

You're right. I should have qualified that statement. Hardware RAID has its place on the enterprise level, but for home use there's just not much of a point. Because my FlexRAID setup depends on individual write speed, I'm only limited by the drives' write speeds or my network and right now they're a little over 140MB/s on average more than fast enough for a home network.

 

[Socio]

So in your use case, if you buy 4 3TB drives and make a 12TB folder (which means you're not using parity), and use a Soft Raid like unRAID, you'll simply keep losing the data on every hard drive that fails. If you make a 9TB array and dedicate one of the 3TB drives as a parity drive, then you'll be able to lose one hard drive in the array without losing any data.

As to hard drives failing, there's all sorts of metrics on this but ultimately, it seems if it doesn't fail immediately, you tend to be fine (for the useful life of your server). There is all sorts of metrics that go into it though and home environments are worse than server environments regardless of what people think. Things that affect a hard drive:

If he wants to feel even safer he could go with a raid 5EE which gives him a built in spare so if one drive fails the spare kicks in until he replaces the failed drive which it will then expand back to the original scheme.

 

[XavierMace]

You're right. I should have qualified that statement. Hardware RAID has its place on the enterprise level, but for home use there's just not much of a point. Because my FlexRAID setup depends on individual write speed, I'm only limited by the drives' write speeds or my network and right now they're a little over 140MB/s on average more than fast enough for a home network.

The catch there is cost. Like I said, you can get that Smart Array for $100 on Amazon right now. unRAID is $70 for up to 7 drives (which I realize is enough for most people) or $120 for more than 7. Given the $30 difference between the basic unRAID and an actual hardware setup, I'd rather just spend the extra for the controller. To each his own I guess.

 

[smitbret]

And if he had 4 or more drives of the same size lying around it would be a fine idea. When half his $200 budget is eaten by the RAID controller and just has a 3TB HDD and a handful of smaller ones then the controller is basically pointless.

With unRAID he could even forgo the license and invest in a 2nd 3TB HDD and possibly another 2TB HDD if he shops around. UnRAID is free up to 3 HDDs and he may never need more than 5TBs of storage. If he does, then he slaps down the $70 for the license and adds any drive he chooses to the array and increases his storage without breaking his array.

The draw of unRAID is the cheap initial outlay for harware, the ability to mix and match drives of different sizes and easily expand the array just by adding whatever drive you find a great deal on without having to worry about copying data and rebuilding arrays.

UnRAID + spare flash drive + Plex plug-in = exactly what he says he's looking for and will work fine with the hardware he already has, leaving his entire budget free for storage.

 

[thecoolnessrune]

Virtualized storage systems have surpassed traditional RAID systems for a long time now. Virtualized storage systems are high performance if designed correctly, and are based on newer technologies that remove the limitations of hardware RAID systems such as:

Limits on hardware. For instance, with ZFS, as long as the system is compatible, you can import an entire ZFS storage array from one operating system to another, as long as the ZFS version is the same or older. Try using another hardware RAID card when one fails and a new one costs a hundred bucks all over again

URE Errors. An era of multi-terabyte drives has made this more common than ever during rebuilds. This is only preventable with scrubs that are crippling to array performance and do nothing about corruption that happens between scrubs and array failures. Modern virtualized raid systems will checksum on the drive to ensure this is caught immediately and corrected.

Write Holes. While a possibility on nearly any RAID system (though hardware RAIDs are not nearly as susceptible), virtualized storage systems use transactional databases to ensure that writes do not get considered complete during a sudden power outage, thus corrupting parity.

Striping. Traditional RAID requires fixed stripe sizes for data writes and rewrites. Depending on the variances in your data this can make performance very bad. Many virtualized storage systems use variable stripe sizes, allowing for more efficient use of disk space and faster transfer speeds on and off the disks.

Tiering. While LSI has their own little CacheCade ability, the vast majority of RAID systems do not utilize auto-tiering storage. Virtualized storage systems do, which provides for much more efficient storage.

Don't get me wrong, traditional RAID still has its benefits. It's simple to set up and soft RAID versions are available in many modern chipsets. But that all said, virtualized storage engines will only become more popular, and is already the native choice for natural storage appliances (systems that run a storage operating system). My guess is that Linux will be the next to catch on. ZFS support gets better by the month, and eventually instead of running md to make a soft RAID, you'll use ZFS in Linux to create a ZFS mirror or RAIDZ setup where you don't have to worry about write holes or UREs. Windows will likely move to ReFS across the board to finally replace NTFS.

 

[smitbret]

Bingo