zoolooemperor

Best way of implementing RAID; Replacing HDDs due to age

Recommended Posts

Hi Everyone!

1) Should hard drives be replaced just because of their age?

I currently use two Samsung Spinpoint F1 HD103UJ 1TB drives, and they're more than five years old. If they are healthy (knock-on-wood), should I start to worry and think of a replace, simply because of their age?

2) Just a couple of days ago, I became a first-time RAID user.

Using the two aforementioned Samsung drives with Windows 8.1 Pro and its "Storage Spaces" feature, I set up a drive pool in a "Two-way Mirror" mode. It should provide a one-disk redundancy similar to RAID1, but it's completely software-based. I decided to go with it because it's easy to set up (apart from having to erase the drives before initializing), and it does not depend on hardware controllers or motherboards. Moreover, and correct me if I'm wrong, it should provide some kind of data integrity checking because of the ReFS file system. The pool is used purely for data storage and access.

Question is: is Storage Spaces recommended as a RAID1-like setup?

There are so many methods nowadays to employ RAID or RAID-like environments, whether hardware: on-board and 3rd party controllers, or software: Storage Spaces, SnapRAID, FlexRAID, DrivePool, Drive Bender (and most likely even more). In your opinion, what is the ultimate?

Please be advised that I'm a Windows user so solutions like Btrfs, ZFS, mdadm and similar are out of the question.

3) What's the best way to set up RAID6 or other two-disk redundancy environment? Storage Spaces for some odd reason requires five drives for such thing, instead of four drives like in the classic RAID6 setup. Plus, my onboard controller does not support RAID6, and it's a pretty advanced motherboard - Asus H97-Pro. So that leaves me with either purchasing a 3rd-party controller, or looking for another software solution.

What would you recommend? Again, we're dealing with a Windows environment.

Thanks in advance.

Edited by zoolooemperor

Share this post


Link to post
Share on other sites

1) I generally operate under the "not broke" policy, but it really depends on how critical the workload is and what the advantages are of new technology in your case. Many worry about using a drive outside of the 3 or 5 year warranty window, just depends on your inclination.

2) Again it depends on the use case and workload but we're generally in favor or RAID1 when it comes to two drives only...we often use RAID10 for larger pools of performance drives.

3) Storage Spaces does allow for single parity as well. Are you wanting to expand the RAID1 you have or set up a new group?

Share this post


Link to post
Share on other sites

1) I generally operate under the "not broke" policy, but it really depends on how critical the workload is and what the advantages are of new technology in your case. Many worry about using a drive outside of the 3 or 5 year warranty window, just depends on your inclination.

2) Again it depends on the use case and workload but we're generally in favor or RAID1 when it comes to two drives only...we often use RAID10 for larger pools of performance drives.

3) Storage Spaces does allow for single parity as well. Are you wanting to expand the RAID1 you have or set up a new group?

1) The data is critical, but I don't think I do any demanding task on those drives. They're just for data storage - data that is accessed constantly, however. Performance is not an issue so I'm not urged to buy newer drives, but I could benefit from purchasing the modern, quieter 5400RPM drives.

2) I was more specifically asking about the best way to implement a one-drive redundancy environment (RAID1 or RAID1-like), not about which RAID is the best.

3) I'm currently with two drives only so I'm limited to a single-drive redundancy, mirroring setup. I might move to two-drive redundant setup in the future, and I'm asking what would be the best way to achieve that, in a Windows environment. I would like to do that with the minimum drive requirement - four - and for that reason, Storage Spaces is irrelevant. It requires FIVE drives for what it calls "Three-way mirror", which is essentially RAID6. I have no idea why!

Thanks.

Share this post


Link to post
Share on other sites

Hi There,

I went through a similar process as yourself not so long ago. I recently purchased a Netgear NAS to provide consolidated, instant access to both critical (documents) and non-critical (media) files. then I also wanted to update my backup solution to backup the data on the NAS.

Firstly, regarding the drive age question:

I agree with the not-broke sentiment above, however I suggest a couple of things to do to reduce your risk of sudden catastrophic failure.

1) pre-clear the disk. This basically just smashes the drive with I/O for an extended period and uncovers bad sectors in the process. (http://lime-technology.com/forum/index.php?topic=2817.0)

2) run something like 'stablebit scanner' on drives that already have data to 1) review SMART data in an understandable way and 2) find bad sectors.

I ended up having to bin 3 drives after these steps cause they weren't up to snuff

Ultimately, I went with FlexRAID for my backup solution for a few reasons which I think may answer your question:

1) It runs on windows and has a GUI

2) it has a number of smart data protection technologies built in (http://snapraid.sourceforge.net/compare.html)

3) It removes some of the drawbacks of keeping backups on proprietary RAID systems, as discussed below;

4) You can use drives that you have lying around to create the backup array and use all their capacity towards the array (either storage or parity protection)

5) It has additional levels of "protection" that are nice compared to something like RAID 5:

  • you can un-fail drives if they get dropped from the array but actually have nothing wrong with them
  • if your array screws up for some reason, you can access the remaining files on the drives using windows, without fancy recovery software
  • Ditto the above if the controller/mobo dies. If this happens with a hardware solution you need to find the exact same controller with exact same firmware if you need to recover the array

I personally consider these things combined as more than 2 levels of protection for my data: 1) Netgear NAS 2)FlexRAID parity to recover 1 failed drive 3) recovery of remaining data in a standard windows environment if it all really hits the fan.

Hope this provides some insight for you.

Edited by AndreasKa

Share this post


Link to post
Share on other sites

Flawless in what sense? How could you be sure of that? SMART values?

Still working, no errors, smart or kernel. Since it's raid, I'll fail one at a time, write zeros at the device level, sector 0 to end run ext smart test , readd, rebuild, and repeat through them all. No issues

Thats how I know. Running zfs too for bitrot issues.

Share this post


Link to post
Share on other sites

Thanks a lot for the replies.

AndreasKa, FlexRAID does sound great.

By the way, your operating environment and needs indeed sound pretty similar to mine. I do have everything backed up, in real-time, but I went for RAID in order to avoid the need to restore in the first place.

1) They offer two products, which one did you go with?

2) What type of RAID did you set up? What redundancy?

3) In what way to you backup to your NAS?

Moreover:

1) Could someone please explain what bitrot is? Is it the same like silent data corruption?

2) How on earth does data become "corrupted"? This just sounds unreasonable. Correct me if I'm wrong, but data doesn't just become corrupted if it isn't caused by a power outages (during drive operations, especially write operations), or a virus.

3) What's an ideal operating temperature for a 7200RPM, 3.5-inch, 1-terabyte SATA HDD?

Thanks!

Share this post


Link to post
Share on other sites

Glad to hear some of my info was helpful.

Answers to your questions below:

1) I'm using the transparent RAID (tRAID) product

2) currently I have 3 drives (2 data + 1 parity) with a view to add another data drive as required. Due to the ability of reading the drives outside of the array if things go awry, I don't feel the need for more parity drives with such a small array

3) I have a windows laptop with a 4 bay enclosure connected to it (the FlexRAID array) and i'm just using SyncBack to backup the NAS shares over the network. Not the most elegant solution but working for now.

Moreover:

1) Bitrot is simply a random data bit being changed 'silently' due to things like the magnetic material losing polarity. It is not detected by most file systems and others use 'scrubbing' (i.e. to re-write the data periodically) to reduce the chance of it happening

2) As above, the amount of magnetic material used to store one bit of information is nowadays incredibly small and the difference between a 1 and a 0 in terms of detected voltage/magnetism is shrinking.

3) most HDDs max temp is stated as 55-60c but if you've ever touched a HDD that's 50+ you know that's already crazy hot. I aim for low-mid 40's

Share this post


Link to post
Share on other sites

I've not used storage spaces so can't comment.

There are many variables but it seems to be that software RAID is the way to go these days from a consumer perspective and nested hardware RAIDs are getting around some of the issues on a larger scale

Share this post


Link to post
Share on other sites

We don't have a stamp for that, lol. But we do use it frequently. Depending on your use case it can be a very effective tool.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now