Best method to store non-critical data? (RAID 5, RAID 6, nothing?)

[sofakng]

What's the general consensus on storing non-critical data like movies and TV shows?

I'm torn between no raid, RAID 5, and RAID 6.

I'm only talking about movies that I've ripped (so I have the physical media as my backup), TV shows, and application/game ISOs (which again, I have the physical media for).

This is data I can afford to lose but of course I'd prefer that I don't.

If I don't go with any RAID and a HD dies than I lose everything on that drive.

If I go with RAID 5 and a HD dies I have the possibility of losing all other drives during a rebuild (eg. second-drive failure).

If I go with RAID 6 I will a LOT of space but am relatively safe. I'm not certain that I need this kind of safety but I'm not sure.

The array size I'm talking about is about 5x 1 TB drives and my non-critical media is currently taking up about 2.5 GB of space.

 

[ss284]

Raid 5 should be fine. Your time is worth more than an extra controller and drive.

 

[a123456]

RAID is okay but if you have that much data that you wouldn't mind losing, a better solution would be an "off-site" backup if you can spare the expense. Something like a separate WHS setup, NAS, or even just a couple of random external HDs that you copy stuff on to once in a while.

RAID 5/6 is good in theory but unless you spend quite a bit to get a quality controller, the rebuilds take forever. It's good for applications where you need 100% uptime (such as business web sites) but might not be as useful in your case since it doesn't sound like uptime is the goal that you're looking for.

Of course, if you really don't care that much about your data, you might be okay with the risks associated with RAID 5.

 

[Swivelguy2]

I agree with a123456. Just use a single drive, but backup periodically to a separate location. It's low tech but protects you in many ways that RAID does not, for the same cost as RAID-1.

 

[seemingly random]

Originally posted by: sofakng
What's the general consensus on storing non-critical data like movies and TV shows?

I'm torn between no raid, RAID 5, and RAID 6.

I'm only talking about movies that I've ripped (so I have the physical media as my backup), TV shows, and application/game ISOs (which again, I have the physical media for).

This is data I can afford to lose but of course I'd prefer that I don't.

If I don't go with any RAID and a HD dies than I lose everything on that drive.

If I go with RAID 5 and a HD dies I have the possibility of losing all other drives during a rebuild (eg. second-drive failure).

If I go with RAID 6 I will a LOT of space but am relatively safe. I'm not certain that I need this kind of safety but I'm not sure.

The array size I'm talking about is about 5x 1 TB drives and my non-critical media is currently taking up about 2.5 GB of space.

Don't know why you would waste the time and money on raid for this case. When's the last time you lost a drive at home without warning.

 

[sofakng]

I've already purchased a Dell Perc 5/i but I kinda regret it and might just sell it off again because of what I learned.

As for backing up, that's not the point. I already have it "backed up" on the physical media so I don't need a 1:1 backup at an external site. This isn't critical data.

 

[a123456]

Yes, it's not critical data but it sounds like you still don't really want to lose it, if for no reason other than the fact that it'll be a giant pain to put in each piece of physical media to restore the data if the HDs die. An external backup would be faster in restoring your data.

If you already have the RAID card, you can set it up. I mean RAID isn't terrible and gets used all the time. As long as you realize the pros and cons and use it knowing the risks, it's fine.

 

[Zensal]

WHS would be a good option i think.

 

[sofakng]

WHS is not bad but if I want to backup everything I lose 50% of my drive space (which seems like a waste).

I'd just go for RAID 5 but it almost seems like it might be worse than ZERO protection. (eg. RAID 5 can cause ALL of my data to be lost but zero protection just means I lose one drive at a time)

 

[pjkenned]

You can turn off duplication for the WHS so you don't lose half your space. I use raid 6 with 10+ 1.5TB drives in my WHS. Plus, WHS does nightly backups which take about 45 sec to set up.

Raid 0 = zero protection. Raid 5 you need to lose two drives before the first one can be replaced and the raid array rebuilt in order to lose everything. At that point you can still read from the array, so if possible, it is always a good opportunity to start pulling the most critical files. Since you have physical media of everything, losing the ISO's wouldn't be the end of the world. Also, with a decent controller, raid 5/6 gives awesome sequential throughput.

 

[RebateMonger]

WHS is a good solution in this case. WHS allows you to use ANY hard drive(s) for storage and makes them all appear as a single drive. You can easily add additional disks to the storage pool, of any size or type, at any time in the future.

No special disk controller is needed, and everything is stored as "standard" Basic disks with "standard" NTFS formatting. If you lose a disk and don't have folder redundancy enabled, you'll only lose the data on that failed disk.

Plus, of course, WHS can automatically make daily full image backups of all the Windows PCs on your network and provide your own web server and remote access server. If you don't want these features (but the backups and backup management are REALLY good and totally automatic), then you don't have to use them.

 

[jimhsu]

Please remember - RAID is not a backup. RAID is a tool that increases redundancy and performance so that failed disks can be replaced.

If you do care at all about your data, use other means of backup - i.e. on an external HDD at a friend's place, to your private FTP server somewhere remotely, to something like Mozy or Backblaze, etc.
If you don't, then have restore plans ready in case of fire, theft, etc. It's not worth it to be paranoid, but it helps to be prepared.

And always weigh the costs of time. If rebuilding, getting a new controller card, etc takes longer than simply re-reading the physical media, then the RAID solution is pretty pointless to begin with.

 

[pjkenned]

Originally posted by: jimhsu
And always weigh the costs of time. If rebuilding, getting a new controller card, etc takes longer than simply re-reading the physical media, then the RAID solution is pretty pointless to begin with.

Plugging a hard drive in takes maybe a few seconds or minutes depending on the drive bays you have. Rebuild max takes 1-2 minutes to initiate (although hot spares go automatically). Reading a full DVD probably takes longer than the user input for the rebuild. In Raid 5/6 your array is still usable after a drive failure so there isn't that much downtime in the whole process. With trayless hot swap backplanes everything can be completed in <3 minutes max, without the server going down.

I would agree that RAID is not a backup, however he also stated that he had physical media, so the storage the OP is talking about is either the backup, or has offline backup already

 

[jimhsu]

Yes, that's the advantage of RAID in that rebuilding is fast. But of course, if the controller dies, then those few days waiting for a new one to ship might be expensive timewise, depending of course on how urgently you need the data. Movies you can probably live without for a few days.

 

[jgigz]

http://www.newegg.com/Product/...2E16859321013&Tpk=acer whs

 

[Madwand1]

Originally posted by: sofakng
The array size I'm talking about is about 5x 1 TB drives and my non-critical media is currently taking up about 2.5 GB of space.

The higher the number of drives, the greater the risk of a single drive failure. Multiple concurrent drive failure is non-zero, but still relatively low.

I'd suggest RAID 5 for a blend of redundancy and storage efficiency. WHS has a high cost / low efficiency for redundancy.

RAID 5 costs you the controller and an extra drive, and some time to select it, learn about it, and set it up. It pays you back in the event of a failure and the additional ripping time at that point.

Note however that if you want easy ability to increase size, then you should look at dedicated controllers with that ability and the largest still-economical drives to start. WHS is superior for expansion due to its simplicity.

 

[pjkenned]

Originally posted by: Madwand1

The higher the number of drives, the greater the risk of a single drive failure. Multiple concurrent drive failure is non-zero, but still relatively low.

I'd suggest RAID 5 for a blend of redundancy and storage efficiency. WHS has a high cost / low efficiency for redundancy.

RAID 5 costs you the controller and an extra drive, and some time to select it, learn about it, and set it up. It pays you back in the event of a failure and the additional ripping time at that point.

Note however that if you want easy ability to increase size, then you should look at dedicated controllers with that ability and the largest still-economical drives to start. WHS is superior for expansion due to its simplicity.

WHS and Raid are not mutually exclusive. For example I have a 12TB usable Raid 6 w/ hotswap array on my WHS that I just partition into 2TB partitions, then use those as drives on WHS. It does take an extra 5 minutes or so to figure out that this is the way to go and then do it, but it is great if you want to rely upon Raid 5/6 instead of WHS duplication.

 

[Madwand1]

Originally posted by: pjkenned
WHS and Raid are not mutually exclusive. For example I have a 12TB usable Raid 6 w/ hotswap array on my WHS that I just partition into 2TB partitions, then use those as drives on WHS. It does take an extra 5 minutes or so to figure out that this is the way to go and then do it, but it is great if you want to rely upon Raid 5/6 instead of WHS duplication.

I was aware that you can use RAID with WHS, but it is typically used instead as an alternative to RAID.

It's interesting that you split up your array into several partitions and manage them separately under WHS -- why did you choose to do that?

 

[pjkenned]

WHS doesn't really like GPT. Made the mistake of making a 8TB GPT volume and had like 2TB usable 6TB unusable. Cuts down on a lot of stress. Further, if a 2TB array gets lost somehow, it is less data. I'd be scared as heck with a 12TB partition. Couldn't even imagine how long it would take to rebuild. Also, and I could be totally wrong here, but it seems like the Adaptec online capacity expansion takes less time with bunches of 2TB drives vs a big drive. No clue why as I tend to only do it when I'm out of town, but just something I noticed.

Plus, I figure easy to buy 2TB drives if I ever need to copy stuff off in an emergency if the raid goes down.

 

[RebateMonger]

Originally posted by: pjkenned
Plus, I figure easy to buy 2TB drives if I ever need to copy stuff off in an emergency if the raid goes down.

As noted, Windows Driver Extender (used to manage disks in WHS) doesn't support GPT at this time. It's an issue for the entire Windows Server lineup, or at least it will be when hard disks larger than 2 TB become available. So there really isn't much choice about partitioning large arrays.

I'm not sure that having 2 TB logical disks under WHS is going to directly help backups, though, since WHS doesn't really distinguish between the different disks. They are all presented as one single storage pool. You'd have to go around the Disk Extender and make disk backups on a lower level than presented by WHS. That may make restoration of the WHS server messy, though. So, while 2 TB backup disks will certainly make backups easier, using them to clone the RAID virtual disks may not be the best way to copy out the data. There's probably a way to do it, but I'm not sure exactly how.

Unless you have high capacity tape magazines, backups of data storage arrays larger than 2 TB is kinda' complicated these days, no matter which version of Windows you use. Or you can send the data to another large storage array.

 

[Madwand1]

Originally posted by: RebateMonger
As noted, Windows Driver Extender (used to manage disks in WHS) doesn't support GPT at this time. It's an issue for the entire Windows Server lineup, or at least it will be when hard disks larger than 2 TB become available.

What do you mean by "issue for the entire Windows Server lineup" when Server 2003 SP1, and 2008 both support GPT? (As do XP-64 and Vista.) Even WHS should support GPT at the low level -- as a simple drive.

 

[pjkenned]

It WHS does "support GPT" in a way. If I recall correctly, if you format a big say 10TB partition as GPT, WHS will use the first 2TB and then the rest of the drive will be considered "system" occupied space when WHS does all of its share location management.

You can get GPT drives to format, mount, and etc no problem. It is the WHS application layer (the Windows Drive Extender as RebateMonger mentioned) that does not like it from what I can tell. Then again, a 12TB partition sounded like a bad idea anyway. It really doesn't make sense to mount a 12TB partition (or even 3TB) if the WHS applications cannot use it.

Not a huge deal, but for about an hour I was trying to figure out how the heck I managed to fill 14TB of space with 2TB of duped data. (I need to work less and sleep more.)

 

[RebateMonger]

Originally posted by: Madwand1
What do you mean by "issue for the entire Windows Server lineup" when Server 2003 SP1, and 2008 both support GPT? (As do XP-64 and Vista.) Even WHS should support GPT at the low level -- as a simple drive.

Things like not being able to back up or restore disks and volumes bigger than 2 TB with the built-in backup tool. And not being able to boot from them.

 

[Madwand1]

Originally posted by: RebateMonger
Things like not being able to back up or restore disks and volumes bigger than 2 TB with the built-in backup tool. And not being able to boot from them.

These are pretty much non-issues for me in the context of Windows servers (and other OS's as described) supporting GPT. If you have say 4 TB of data, and enough space somewhere to back it up, I think you can figure out the backup tool without getting stuck on the limitations of the free built-in versions. Similarly if you have 4 TB data in an array, and you're trying to use the same array for your OS, I think you're just doing it wrongly.

Granted that WHS apparently has issues with large arrays, but I think I'll just assume that I read your statement "...doesn't support GPT at this time. It's an issue for the entire Windows Server lineup" too broadly and call it a day.