Educate a noob on redundancy

[GodlessAstronomer]

I'm building a new system very soon and one of my biggest priorities is data backup/redundancy. Since I started using computers 15 years ago my backup strategy has been beyond pathetic, a mish-mash of burning disks and shifting data to extra (usually older) drives whenever the idea struck me (not very often). I've lost some data over the years, but miraculously not a lot.

Well that's all going to change with my new system and I'd like your help. What's the best way to ensure my data will not be lost due to HDD failure? I really know very little about RAID or how it works. I'm currently looking at getting 2x 1tb Samsung Spinpoint drives, what's the best way of making sure my data is backed up? Should I just copy over an image of one drive to the other every week, or should I have some kind of RAID setup?

The way I think right now is that I'll have a 1tb drive that is partitioned with 200gb for my OS and the remainder for data. Is it possible to run this setup in RAID without having to buy a separate RAID controller? Also I know there are a bunch of different kinds of RAID. Which is best suited to my application? I don't mind losing half of my storage capacity. Speed is a priority, but data integrity is the number 1 priority.

I don't mind if you just post links to articles, I know RAID is too big of a subject to discuss at length here, but Googling the subject brings up a lot of redundant or outdated results so it's hard to know what to read.

 

[her209]

What if your place of residence goes up in fire?

 

[GodlessAstronomer]

Originally posted by: her209
What if your place of residence goes up in fire?

Unfortunately bandwidth in my country is absurdly expensive so I can't back up my music, video, applications and OS stuff online, but I have online backups of all of my important documents and projects. I'm asking about local backup here.

 

[Zap]

Originally posted by: GodlessAstronomer
What's the best way to ensure my data will not be lost due to HDD failure?

Do you care at all about possible data loss from accidental deletion, file system corruption, viruses and other people screwing with your system? Just understand that there are a lot more ways to lose data than from actual hardware failure of the HDD.

Do you care at all about how fast it will be to get your system back up and running?

For just not losing data, you want more than one copy.

For quickly restoring your system to an operational state, either use RAID1/5/6 with a hardware controller that can rebuild on-the-fly, or use a backup software with an external HDD that does partition imaging (like Acronis, free if you own a Seagate or WD drive).

For just not losing data, either manually make copies to an external drive or use backup software to copy for you to an external drive. Backup software can manage copies by automatically copying only what is needed (new or changed files).

 

[aE0n]

The best thing to do is get an external drive or second internal drive and configure automatic backups. RAID is not a backup solution but a data reliability solution.

For example, if a drive dies in a RAID mirror, you can still work with no impact. But if your files become corrupt they will be corrupt on both drives in your RAID mirror.

You can RAID your data, but a backup is still recommended.

 

[elconejito]

As it is so often said around here... RAID is not a backup.

The first thing I would recommend is using an external backup drive.
-If your computer explodes, you will still have the external drive.
-If you accidentally delete a file from your internal drive, you still have a copy on the external
-If your house burns down, it's easier to just grab the external and run than to carry out a whole desktop.

If you're really paranoid (like I am) then you need to worry about what if something happens while you're not home. So you need an "off-site" backup. If online backup solutions are too expensive, then get another external drive and keep it at someone else's place like parents, siblings, uncles, friends, etc. You can even make a deal where you keep their data and they keep yours.

RAID makes it easy to recover if a drive goes bad... immediately (or at least very quickly). In RAID1 if one drive fails you can still keep working on the mirrored drive. In RAID5 or RAID6 you replace the damaged drive, rebuild the array and your back in business. So if reduced downtime is what you are looking for, then RAID is a great solution... but it's still not a backup 🙂

 

[GodlessAstronomer]

Originally posted by: Zap
Do you care at all about possible data loss from accidental deletion,

yes

file system corruption,

yes

viruses

yes

and other people screwing with your system?

No (not going to happen)

Do you care at all about how fast it will be to get your system back up and running?

Not a priority

Thanks for the help guys. Sounds to me like RAID isn't the appropriate solution. I'm thinking 2x1tb drives, one internal and one on e-SATA. Set up an automatic backup every night to copy important stuff over to the e-SATA drive. Does this sound about right? Any other suggestions?

I didn't want to get into it here but seeing as some of you have brought it up, regarding off-site backups. My most important data (RAW and JPG photos, documents and projects/assignments) is backed up on my laptop and on burned DVDs. Videos and music I'm not too worried about - it would suck to lose all of that but it's not critical (but will be backed up on the e-SATA).

 

[Emulex]

raid is good for single drive failures if you use high quality raid drives. otherwise it is miserable.

backup your system(s) to a storage server then back that up to external disk (or three).

if your systems get infected you want to make sure you have enough off-site storage that you can rotate that copy back in. One drive would leave a point of failure since the infection could occur right before backup. or the drive could fail when erasing/writing the new backups.

 

[RebateMonger]

When offsite backups are difficult or impractical, it makes sense to make thorough local backups and back up what you can afford online. If you have relationships with other computer owners, there's also the option of exchanging backup disks or sending backups across the Internet or VPN to a PC located at somebody else's place. You can encrypt the data before sending it.

The simplest, most automatic, and most complete backup system I'm aware of is Microsoft Windows Home Server. It'll only work with Windows XP and later. WHS automatically makes full image backups of all Windows PCs on your network. It automatically manages those backups to minimize disk space and retains multiple versions of every file, going back months. If you delete or corrupt a file, you can restore that file (or restore the whole PC) to the way it was yesterday or six months ago.

Note I'm NOT talking about using WHS as a file server. I'm talking about using it for making backups. If you use it as a file server, holding your ONLY copy of data, you'll want to back IT up, just like any other file server. WHS has built-in folder redundancy available, but redundancy, whether it's WHS' form or a redundant RAID system, doesn't solve the "Drats. I accidentally deleted 1000 files" problem. Backups solve this problem. Disk redundancy doesn't.

WHS also has extremely simple disk management. If you need additional storage, you drop or plug in another disk (any size, any type), do a couple of mouse clicks, and the extra capacity is added to your storage pool.

 

[GodlessAstronomer]

There's no way I can fit anything even close to that in my budget RebateMonger. This isn't the first time you've tried to shove WHS down my throat when it's quite obviously way out of my budget. Please stop.

 

[pjkenned]

If you just want something simple, either use RAID 1 (mirror the 1TB drives) or just install them both and use one as a data drive, one as a place to do windows backups to. Windows 7 and Windows Vista can do automatic backups to an internal drive. Raid 1 has the nice feature of not giving you any lost time. Backup to a second internal drive gives you some safety if you change something on your drive that screws up everything (i.e. you delete 1000 files as RebateMonger suggests).

 

[RebateMonger]

Originally posted by: GodlessAstronomer
There's no way I can fit anything even close to that in my budget RebateMonger. This isn't the first time you've tried to shove WHS down my throat when it's quite obviously way out of my budget. Please stop.

Sorry, but I don't remember your post. I answer a LOT of questions about a lot of topics. In this post, you didn't indicate you'd posted on this topic before and you didn't mention your budget. You said your backup strategy is "pathetic". Like most of us, if whatever backup you choose isn't automatic, it won't be used.

If you only have one PC, then attach a large second disk (internal or external), leave it attached, and set up an automated backup program to write to it often enough that you keep the data current but seldom enough that you'll be able to recognize that something's wrong and can still recover any accidentally-deleted data.

Or use an archival backup where the entire system is backed up infrequently and only the changed data is backed up frequently. It's pretty space-efficient. Just test it to be sure it's working and that you are backing up what you intended to back up. That only costs whatever a second disk costs, plus the cost of any needed backup software.