What are you doing with your HDDs?

[Michael73]

For all those upgrading to the new Mac Pro (from an existing one), what are you doing with your existing drives?

For my part, after spending ton of time researching and getting advice from the business folks at the Apple Store, I'm getting a Drobo 5D to hold all my media files and for my Time Machine backups.

 

[AidenShaw]

For all those upgrading to the new Mac Pro (from an existing one), what are you doing with your existing drives?

For my part, after spending ton of time researching and getting advice from the business folks at the Apple Store, I'm getting a Drobo 5D to hold all my media files and for my Time Machine backups.

If your drives are more than a couple of years old, seriously consider buying new drives. At the very least, run a full S.M.A.R.T scan and toss any drive with bad blocks (re-allocated blocks) or other errors.

And, don't get standard desktop drives for a RAID (or Drobo). Pay a little more and get enterprise or NAS drives - they're made to run 24x7 and play nicely with RAID controllers. (I have about 20 of the Seagate ST4000VN000 4TB NAS drives and have had no problems. If you're in the Bay area, they're $199 at Central...)

 

[Michael73]

If your drives are more than a couple of years old, seriously consider buying new drives.

I guess, but if you're sitting on top of a pile of drives which almost anyone who upgrades to the nMP will be, doesn't it make sense to use the ones you have and then replace them sequentially as they fail? I'm mean your data is protected as it's in RAID.

Put another way, provided your drives are working today, wouldn't you rather save the $1,000 you'd spend on 5 x 4TB Seagate drives and put that money towards a higher spec nMP?

 

[DanielCoffey]

If your data is business critical, which it probably is for the intended user of the nMP, it will make sense to replace drives *before* they fail. Downtime is lost money.

Check SMART and chuck any that are ageing.

 

[slughead]

Put another way, provided your drives are working today, wouldn't you rather save the $1,000 you'd spend on 5 x 4TB Seagate drives and put that money towards a higher spec nMP?

If you have a RAID01 or RAID10, this is the whole point.

Even consumer hard drives often last a half million hours or so, provided you have good clean power etc. I would replace them as they fail, then rebuild when the system is idle.

 

[Studio K]

At the very least, run a full S.M.A.R.T scan and toss any drive with bad blocks (re-allocated blocks) or other errors.

What is a good software to throroughly check a drive for bad blocks? MAc or Windows software.

 

[Michael73]

If your data is business critical, which it probably is for the intended user of the nMP, it will make sense to replace drives *before* they fail. Downtime is lost money.

Check SMART and chuck any that are ageing.

Downtime *is* lost money but spending money on things that aren't broken can be wasting money as well. There needs to be a balance.

The fact that your data is in RAID means there is no downtime except for that which is associated with going to buy a new drive.

Here's another thought…drive prices invariably fall over time so while you may be buying yourself more time by getting a new drive up front you're probably spending more money than if you'd waited.

 

[AidenShaw]

I guess, but if you're sitting on top of a pile of drives which almost anyone who upgrades to the nMP will be, doesn't it make sense to use the ones you have and then replace them sequentially as they fail? I'm mean your data is protected as it's in RAID.

If they aren't "Enterprise" or "NAS" or "RAID Edition" drives designed for a RAID array, I would never use them in a RAID-5/6/50/60 array.

Desktop Drives Drop In and Out

Desktop Vs Raid Edition drives

Using desktop drives under the RAID environment could result in frequent drives being dropped out of the RAID group while trying to recover bad sectors during read operations.

Enterprise drives uses a special FW for error recovery that limits the time it takes to attempt to recover a bad sector as other drives are waiting during that process.

Desktop drives are designed in scenarios where there is no data redundancy or other drives working in a group so it has to spend additional time to attempt to recover data.

http://kb.promise.com/Print10224.aspx

This isn't theory.

Put another way, provided your drives are working today, wouldn't you rather save the $1,000 you'd spend on 5 x 4TB Seagate drives and put that money towards a higher spec nMP?

No, I wouldn't and I don't. Why spend all that money on a rig and cheap out on the storage and risk losing your data? Maybe use them for a big backup volume, but not for data that you wouldn't want to lose.

If your data is business critical, which it probably is for the intended user of the nMP, it will make sense to replace drives *before* they fail. Downtime is lost money.

Check SMART and chuck any that are ageing.

Good advice.

What is a good software to throroughly check a drive for bad blocks? MAc or Windows software.

"Hiren" has a freeware bootable CD with a number of Linux tools on it, including partition management software and a good S.M.A.R.T. utility.

http://www.hirensbootcd.org/download/

The fact that your data is in RAID means there is no downtime except for that which is associated with going to buy a new drive.

Wait until you lose 10 TB of data because a second drive failed during the rebuild - which is actually fairly common because the stress of a rebuild can put a weak drive over the edge.

Saying that "RAID means there is no downtime" is a complete misconception.

There's an adage that the paranoid wear both "belts and suspenders". With disks, you need to wear belts, suspenders and underwear - so that if two of the three fail your junk isn't waving in the breeze.
__________

I'd keep older drives if *all* of the following are true:

• S.M.A.R.T. shows no re-allocated sectors or "pre-failure" or "old age" warnings
• The drives are all "RAID Edition" drives or the equivalent
• You set them up with RAID-6 or RAID-60
• You have hot spares that will automatically rebuild the moment a drive fails

Drive failure is not a probability, it is a certainty. Sometimes you're lucky, sometimes you're not. If you don't have hot spares, the "unprotected" window is "until you notice". If you don't use RAID-6/60, the chance of losing the array to multiple failures is too high, even with hot spares.

I have a system with 180 drives. If my drives have a 3 year MTBF, I'll have 3 to 4 drives fail per month. And they do - so I order spares by the 20-pack.

 

[Varmann]

Even consumer hard drives often last a half million hours or so, provided you have good clean power etc. I would replace them as they fail, then rebuild when the system is idle.

Do not take the MTBF figures literary, it is not a measure of average life span.
No drive lasts 100 years.

 

[AidenShaw]

For my part, after spending ton of time researching and getting advice from the business folks at the Apple Store, I'm getting a Drobo 5D to hold all my media files and for my Time Machine backups.

Separate post, for a separate thought.

If your disks are different sizes, Drobo will choose "not to use" some of the mismatched space. (I avoided saying "waste" - there's an obvious technical reason why they do this.)

If you put a 4000GB disk and a 100 GB disk into a Drobo, you'll get 100 GB of storage.

Check the Drobo website - they have a pretty clear explanation of the algorithm used when you have multiple disks of different sizes.

----------

Do not take the MTBF figures literary, it is not a measure of average life span.
No drive lasts 100 years.

Actually, in spite of my recent post - I do have a system that I set up in 2004 still running with 9 year old disks.

The disks are silly expensive enterprise SCSI, with crazy big cache RAID controllers and RAID-60 and multiple hot spares.

...and I do an incremental backup every 12 hours.

Some "5 year MTBF" disks may last for 20 years, and others will fail the first week. That's what "mean" means.

 

[ActionableMango]

The fact that your data is in RAID means there is no downtime except for that which is associated with going to buy a new drive.

Hahahaha.... no downtime? That's funny!

Have you thought about what happens if the Drobo fails? No? Perhaps you should read their support forums.

RAID protects you from hard drive failure, and that's it. It doesn't protect you from a myriad of other problems. You've just changed the single point of failure from the drive to the RAID enclosure.

 

[AidenShaw]

Hahahaha.... no downtime? That's funny!

Have you thought about what happens if the Drobo fails? No? Perhaps you should read their support forums.

RAID protects you from hard drive failure, and that's it. It doesn't protect you from a myriad of other problems. You've just changed the single point of failure from the drive to the RAID enclosure.

And don't forget about wetware failures - if you're told to replace "drive 3" and you don't know if the controller calls the first drive "drive 0" or "drive 1" you could lose everything in an instant.

Our "fools" get better every month - "fool-proof" doesn't exist.

 

[alphaod]

Seeing my current setup is a 512GB SSD, and a RAID 5 of 4x 300GB 10k drives (net 900GB), I think I will survive with a 1TB SSD and perhaps an external Thunderbolt enclosure. I already have a Drobo which runs off USB 3.0; I also have a Promise Pegasus J4.

 

[Varmann]

Separate post, for a separate thought.

Actually, in spite of my recent post - I do have a system that I set up in 2004 still running with 9 year old disks.

… snip ...

Some "5 year MTBF" disks may last for 20 years, and others will fail the first week. That's what "mean" means.

Exactly!
But MTBF figures today is in the order of 1 million hours for enterprise disks (thats more than 100 years). This is not the expected life span of a single disk but related to how many of, say 10,000 disks, that is expected to fail the first month or so. No disks lasts 100 years, 10, maybe 20 years if you are lucky, but not much more than that. It is a mechanical product and it has parts that will age, it is just a matter of time before it finally fails.

 

[Michael73]

Hahahaha.... no downtime? That's funny!

Have you thought about what happens if the Drobo fails? No? Perhaps you should read their support forums.

RAID protects you from hard drive failure, and that's it. It doesn't protect you from a myriad of other problems. You've just changed the single point of failure from the drive to the RAID enclosure.

Did you not read my original question? It asks what you're doing with your existing drives? Any single system can fail. I never said the Drobo would be my only backup. That's why I utilize off-site backup as well. After all, what happens if my house is broken into and the Drobo were stollen? What happens if my house were struck by lightening (like in 2006) and fried all my electronic components? Or, a hundred more scenarios.

And don't forget about wetware failures - if you're told to replace "drive 3" and you don't know if the controller calls the first drive "drive 0" or "drive 1" you could lose everything in an instant.

Our "fools" get better every month - "fool-proof" doesn't exist.

Drobos have lights on the front next to each drive that show which one has failed. I shouldn't need to figure it out by volume or drive name.

 

[Varmann]

Did you not read my original question? It asks what you're doing with your existing drives? Any single system can fail. I never said the Drobo would be my only backup. That's why I utilize off-site backup as well. After all, what happens if my house is broken into and the Drobo were stollen? What happens if my house were struck by lightening (like in 2006) and fried all my electronic components? Or, a hundred more scenarios.

I will make "yet another full backup".
I have 5 complete backups so far, two I rotate offsite + a couple more of data I can´t reproduce. I am a bit paranoid but it has saved me from losing any data since -91 (when I almost lost all data for my PhD, luckily I could reproduce it).
A drobo sounds like a decent extra backup to me.

 

[Rich.Cohen]

Michael,

I'm in about the same boat you are in. I have a very old Mac Pro, but most of my Hard Disks are less than a year old. I'd like to keep them although I'm prepared to buy some new disks as well. I'd also like to avoid transferring data from my old disks to a new setup. I expect data transfer to be a chore since my current Mac Pro is a pre 2008 model 1,1.

My ideal would be a RAID enclosure that would allow me to insert two existing 3 TB drives with data and two new drives and then let the RAID duplicate the data. I have no idea how practical that is and I haven't yet found an enclosure that will fit the bill.

Rich

 

[Michael73]

Michael,

My ideal would be a RAID enclosure that would allow me to insert two existing 3 TB drives with data and two new drives and then let the RAID duplicate the data. I have no idea how practical that is and I haven't yet found an enclosure that will fit the bill.

Rich

One thing to note about the Drobo is that it will reformat any drive you put into it so you can't just take a drive out of your MP and plop it into the Drobo. If I was in your shoes I would take the two new 3TB drive and put them into the Drobo and let it format them. According to the size calculator on their website that gives you 2.72TB available for data, 2.73TB used for protection and 4.88GB for overhead. I would attach the Drobo to your existing MP and copy the entire contents of ONE of your 3 TB drives. Once completed, take the original drive you just copied from out of the MP and put it into the Drobo and let the Drobo format that. Now, you'll have a total of 5.44TB available for data, 2.73TB used for protection and 9.77GB used for overhead. Once again, copy the contents of the second 3TB drive still in the MP to the Drobo. Finally, when finished take the drive out of the MP and put that into the Drobo and let the Drobo do it's thing. At the end of the day, you'll wind up with 8.17TB available for data, 2.74TB for protection and 14.66GB for overhead.

 

[TwoBytes]

What are you doing with your HDDs?

Even raid 6 may be at risk! :/

http://www.zdnet.com/blog/storage/why-raid-6-stops-working-in-2019/805

By then we hope to have SSD arrays if they are cheap, have large capacities and are quicker to rebuild.

I'm that paranoid to have two drobos. No downtime and two copies.

 

[haravikk]

Whether I end up going for the Mac Mini or Mac Pro I'll probably just settle for a basic USB four bay enclosure and rely on internal SSD to create a Fusion Drive. I've been looking at loads of fancier options, Thunderbolt in particular, but none of the options are desperately affordable. The Drobo Mini is very enticing, but it's one that I would need to buy all new drives for, so it's most likely to also end up being a future upgrade rather than an initial update.

That said, I've been investigating some more unusual options as well, such as self-building an 8+ bay enclosure using either a regular PC case or a DVD Duplicator case, this way I can consolidate all my drives into a single enclosure instead of several, though I'd probably have to find somewhere out of the way that I can put it that isn't so far its performance will suck.

There are 8-bay and more enclosures out there, but the Thunderbolt ones are insanely expensive, and the others require eSATA or Mini SAS, or have limited to poor support for USB 3 on Mac; plus the cheap ones aren't so cheap that building your own couldn't be a much better option even though it may work out as a bit more overall.


But yeah, the short answer would be "I don't know yet, but probably whatever's cheapest to start with with a view to upgrading later".

 

[deconstruct60]

Even raid 6 may be at risk! :/
...
By then we hope to have SSD arrays if they are cheap, have large capacities and are quicker to rebuild.

But SSDs fail also. Especially, if you go cheap. Fact is most SSDs have a layer of something like RAID-5 ( single parity protection) in them already. Read bit failures happen from them also even with their protections from media collapse failures.

 

[deconstruct60]

I'm mean your data is protected as it's in RAID.

If there is a disk failure it means your data was being protected. In diminished capacity mode there is no protection. If the window on second failure happens to coincide with your first failure you are toast.

This is why proactively pull drives before they completely catastrophically die. Done properly and are far more likely to widen the window between individual failures and hence increase safety.

Put another way, provided your drives are working today, wouldn't you rather save ...

There is a difference between appears to work and working. If the drive is giving indications of upcoming failure then it isn't really "working". For example, remapping sectors because they can't hold data correctly isn't odd description for working.

If the drives are being moved from one system to another anyway, there are few better times to do deep diagnostics to test the health of the drives. A decent size set of drives that is several years old that have been in constant use has a relatively good chance of turning up some drive that is questionable.

 

[TwoBytes]

This is why I like drobo. From what I told, it periodically tests each drive and if a drive fails the criteria (similar to SMART) it marks the drive as bad and asks you to replace it. It's a little bit more of an early warning than just a failed drive

 

[AidenShaw]

If there is a disk failure it means your data was being protected. In diminished capacity mode there is no protection. If the window on second failure happens to coincide with your first failure you are toast.

This is why proactively pull drives before they completely catastrophically die. Done properly and are far more likely to widen the window between individual failures and hence increase safety.



There is a difference between appears to work and working. If the drive is giving indications of upcoming failure then it isn't really "working". For example, remapping sectors because they can't hold data correctly isn't odd description for working.

If the drives are being moved from one system to another anyway, there are few better times to do deep diagnostics to test the health of the drives. A decent size set of drives that is several years old that have been in constant use has a relatively good chance of turning up some drive that is questionable.

I call this a very good post. I tip my hat.

Be proactive about replacing drives - don't spend thousands on a new Apple and lose all of your data because you didn't want to spend a couple of hundred more on new server class drives.

And don't buy the cheapest desktop drives for a multi-drive RAID setup. Buy NAS/RAIDedition/Enterprise drives. They're designed for 24x7 operation, they have firmware to play nice with RAID controllers, and they're designed to deal with the horrible vibration issues that happen with multiple drives tightly coupled in a small chassis.

 

[Rich.Cohen]

Finally, when finished take the drive out of the MP and put that into the Drobo and let the Drobo do it's thing. At the end of the day, you'll wind up with 8.17TB available for data, 2.74TB for protection and 14.66GB for overhead.

Thanks. That sounds like a very workable approach to my needs.

I just read through reviews on Amazon. There are some reports of total failure of the Drobo unit. In a few cases this resulted in total loss of data. These cases seem to be rare, but it brings up the question of how to best protect a large quantity of data. I have about 7.5 TB of data. At present, the most critical data (about 1 TB) is backup using TimeMachine to a 3TB external HD. The remaining data which is very stable (few deletions, no updates and new files added periodically) is backed up to two external HD using chronosync.

I'm going to buy a new Mac Pro when it comes out. A drobo with 2 4TB and 3 3TB drives will give me some room for room for growth and protect me from a single disk failure. I can also keep my current TimeMachine drive. If I want to increase my level of protection, what are my options? Buying a second Drobo with another 17TB of storage seems like overkill. Is there a middle ground?