Any recommended External backup drive for backups ?

[johnpau80]

We are planning to move to the external hard drives for backups instead of the slow tapes.


Any recommended industrial strength external hard drives, as I had seen in some of the maxtor external drives, during swapping or removing, these hard drives keeps throwing errors unable to read or disk not rewritable..etc..

 

[SjrH]

Looked into this last year but the problem was hot swapping drives over.

You could always look into Iomegas REV drives - there a lot faster than tapes but just as easy to swap around

 

[riverider]

My choice is one Mfg. by Acomdata they are sold by CompUSA and they are a quality drive.

 

[kmcferrin]

I definitely wouldn't recommend the REV drives, or any other proprietary media format. For one thing, they tend to be more expensive than standard media (LTO tapes, IDE hard disks, etc). They also tend to come and go with the years. Before REV Iomega also had ZIP and JAZ drives for storage, and while they were especially prevalent at one point you can't even find them on ebay these days, which could make restoring an old backup difficult if your drive fails in the future.

I would say stick with IDE/SATA had disks in a hot-swap chassis if you must use removable hard disks for backups.

 

[kjv1611]

It wouldn't be a bad idea to use a RAID configuration, as well as keeping a tape backup if your data is valuable. After all, the costs of the backups would be nowhere near as much as the costs of retrieval if you had a complete hardware failure which is unlikely, but still possible. For instance, many businesses now use RAID configurations for local backups, but still also use tape backups that get stored at external locations (mostly for crisis management - such as what if our entire building gets burned down or blown up).

 

[tf1]

Use an HP DL380 (or similar) Data Protection Storage Server running Microsoft System Centre Data Protection Manager 2006 (DPM). This provides near continuous protect taking hourly rapid snapshots of your data with blindingly fast restoration times in event of a system failure.

Regards: tf1

 

[Auger282]

with DPM you still need a tape backup for long term storage....

The tapes arent slow.. your drive is slow.. LTO's the way to go... DPM servers are only if you have big big file servers, multi file servers, stupid users that always need restores.

We use an undelete program on our file server with an LTO ultrium 2 drive (320) and it works fine for us (150 gb file server).. our file server has raid 5..

Raid is for hardware redundancy not data.. thats what backups are for...

 

[tf1]

Much depends on the requirements of your backups. If you are running data that is mission critical, with hourly snapshots DPM ensures that losses are minimal and that restoration is maximised.

You still need tape backup and I agree that Ultrium/LTOs are very fast. They need to be connected to an exclusive SCSI channel, the faster LTO3s need an U320 channel for max performance.


Regards: tf1

 

[kjv1611]

Here is a definition of redundancy from

http://www.dictionary.com:

6. Electronics. Duplication or repetition of elements in electronic equipment to provide alternative functional channels in case of failure.
7. Repetition of parts or all of a message to circumvent transmission errors.

Those were copied ver batim. So, yes, RAID is for redundancy, which in itself is a "backup" of the data, system, whatever. Everything else is extra backup.



Bottom line - these days, there are many options for backup - it just depends upon what your and your company's preferences are.

 

[kmcferrin]

You can quote the dictionary if you like, but for most business purposes (and certainly the legal requirements of laws like HIPAA and Sarbannes-Oxley) RAID isn't a backup. All RAID does is protect you against a disk failure. If someone intentionally deletes data that you need, you cannot restore it from a backup if you're only using RAID. If a file becomes corrupted, you cannot restore it from a backup if you're only using RAID. If your server catches on fire, your building is destroyed, or even something as simple as your server is stolen, you cannot restore it from a backup if you're only using RAID.

RAID is for fault tolerance. It isn't a backup solution. If the businesses in the World Trade Center were relying on RAID for a backup solution instead of proper, off-site, tape backups then far fewer of them would have recovered. The same goes for the businesses hit by hurricane Katrina. According to the Red Cross, 60% of established businesses that do not resume operations within a week of a disaster occurrance do not come back. Even still, of those that come back 40% of those go out of business within 2 years. If you do the math, that means that on average only 24% of businesses affected by a disaster actually are able to resume long-term operations. With numbers like that, why on earth wouldn't you want to actually protect your data properly?

 

[tf1]

I agree entirely. RAID isn't a backup. It is either for performance (RAID 0), disk fault tolerance (RAID 1) or performance and disk fault tolerance (RAID 1/0, RAID5 or RAID 6).

At least the latest backup set needs to be stored off site in event of a disaster, theft or vandalism. Preferably, all backup sets should be taken off site except the set scheduled to run.


Regards: tf1

 

[gbaughma]

kmcferrin makes some excellent points.

I work in a situation where we have to be both HIPAA and DOC (Department of Corrections) compliant.

RAID will keep your system up in the event of a hard drive failure (if properly implemented; striping and hot-swappable)... note I said "A" hard drive failure. If you lose two at the same time, you're in a world of hurt (referring to Raid 5)

However, every company should have a disaster recovery management (DRM) plan.

Here's how ours works.

I currently have 14 servers in my server room; we migrated to about 90% thin clients about a year ago. So no user files are stored on local machines; they're stored on servers.

Servers all have redundancy; RAID 5, redundant power supplies, UPS capable of keeping all servers running, and tied in with UPS management software so that when the batteries reach a critical drain point, the servers to a graceful shutdown, instead of just losing power.

We have a Dell PowerVault DLT-4 tape library, 8 slot. The 8th slot is loaded with a cleaning tape. We have 4 sets of 7 tapes in each set. We run a full backup every night (not incremental... complete). Each Tuesday morning, the tapes are rotated; the oldest tapes taken from a fire-proof filecabinet and inserted; the tapes that were just removed from the library go to a bank where they are put in a safe deposit box.

GHOST images of all the servers are done every 6 months, written to DVD-R's, and put in the bank vault.

Every month, an additional "monthly backup" is run; a complete backup of all servers on a separate blank tape, which is labeled and taken home with me. When I take a new monthly home, I bring the old monthly back, which is put in the fire-proof cabinet. This way, the most recent daily backups are in a safe-deposit box offsite, and the most recent monthly backup is stored offsite at my residence.

At the completion of every tape backup, the software fires off an e-mail to me telling me if the backup was a success.

Tape overwrite on daily backups is set to 3 weeks, since we're on a monthly rotation.

Critical systems, such as our intranet web server information is backed up on CD, so that it can quickly be deployed onto another server if necessary.

Users understand that if they *do* have a PC or laptop, that there are no backups done of their system. If they feel something is valuable, it is to go on their user share.

We have images for every type of computer on a dedicated GHOST server. Those images are also at our corporate office 1200 miles away.

In the event of hardware failure, we have a 4-hour contract with the manufacturers. I had a new SCSI drive for one of our RAID arrays in my hands in 4 hours 10 minutes, which isn't bad considering how far out in the boonies we really are.

RAID is not a backup. It's a redundancy. Ask yourself this question, when thinking about your DRM Plan... if your server room flooded from a busted sprinkler head, or caught fire, how long would it take you to get up and operational again? Could you be limping along with data restored by the next day? Same day? Within 7 days? And how much of an impact would that have on your business? Would your business even survive the amount of downtime?



Just my 2¢

"In order to start solving a problem, one must first identify its owner." --Me
--Greg

 

[kjv1611]

Note to self: Dictionary bad.

Seriously, though, kmcferrin, no harm was intended with the quote from a dictionary. Sorry if any was taken.

 

[LWComputingMVP]

This question specifically asks about what drives for the backup solution. Personally, I prefer the ComboDock from

http://www.wiebetech.com

as it allows you to use ANY internal IDE drive for backups and can be swapped out while the system is running. I use these at a couple of clients.

There has been some debate over REV drives and hard drives and RAID and such -

http://www.lwcomputing.com/tips/static/backup.asp

contains my take on these topics - I wrote it to address the numerous backup questions I've answered on other sites. Note: outside of a couple of instances (related mostly to REV drives and Wiebetech) I do not recommend specific products. This is a conceptual overview.

 

[gbaughma]

You know, I was reading the other day (on TT) that a government study shows that the average life of CDRW/DVDRW discs is about 3 years.

So much for that "Good for 100 years" crap they were hyping to begin with.

Then I started asking myself (for those of you who wondered how this ties in with the thread) about the life of hard drives.

According to Smart Computing

http://www.smartcomputing.com/edito...rticles/hardware/2004/h0707/harddrivesbuy.asp

the MTBF on a hard drive is 400,000 to 1.2M Hours.

At 400,000 hours, that means that *ON AVERAGE* a hard drive should fail every 45 YEARS. At 1.2M Hours, that's an average failure of about 137 years.

Well... **** WAIT A MINUTE **** ... 45 YEARS? Has anybody here ever had a hard drive last 10 years? Where are they getting these numbers, of 400,000-1.2M MTBF? Or have I just had EXTREMELY bad luck, and I keep getting the drives on the low-end of the curve, that only last 3-5 years, instead of an average of 45?



Just my 2¢

"In order to start solving a problem, one must first identify its owner." --Me
--Greg