Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

prvt.donut

macrumors 6502a
Original poster
Jan 1, 2008
525
26
A recent post by yet another user whose RAID 5 array failed and due to unautomated backups is now facing a huge bill to recover missing data from a data recovery company that led me to create this thread. Too many companies and IT professionals haven't had this drilled into them enough and as a consequence will face dire situation where they face potential data loss.

Ask yourself this question. "Can you afford to lose your data?"

Data integrity is of paramount importance.​

It is vital that if you rely on and need your data for your livelihood that it is adequately backed up and off-site replicated.
Here is the holy trinity of data protection, as long as your plan encompasses these basic points, you should be prepared to recover from a critical data loss:
  1. Daily reliable backups.
  2. Off-site replication.
  3. Disaster recovery.
If you plan these 3 items then you can be sure of your data integrity and your business workflows will be protected. These can be achieved at nearly any price point.

I would hope that by compiling knowledge from Mac Professionals this can become a resource for all professionals to ensure they avoid a huge bill and the ulcers associated with trying to deal with when your RAID arrays fail and are facing data loss of work that you have spent precious time working on.

So, for the Mac professionals out there. How do you handle your storage, backups and prepare for disaster recovery?
 
Last edited:
Never use RAID for backup, except simple mirroring. RAID is not a backup solution, well know by pros. RAID may be useful for performance purposes or to keep you going during recovery.
 
  • Like
Reactions: mrxak
Never use RAID for backup, except simple mirroring. RAID is not a backup solution, well know by pros. RAID may be useful for performance purposes or to keep you going during recovery.

Simple mirroring is not an exception - RAID-1 is not a backup.

On the other hand, parity RAID or mirroring (RAID level 1,10,5,50,6,60) might help you to never need your backups - so RAID can be a very important part of the overall strategy.

RAID can not protect from the human "Ooops!" factor, nor from malware like ransomware. It can turn an event like a failed drive from an extended downtime to a minor annoyance.

Parity RAID (or similar replication) is also an important consideration for the backup server. You don't want to find out that your single disk backup device has a bad drive when you go to recover lost data.
 
Aren't you being overly hard on the guy? It sounds as if he was doing what he could within the constraints he had. He had a RAID5 configuration and was performing some level of backups. He just had a bit of bad luck in that he had multiple drive failures at the same time in his RAID5 configuration.

While enterprise backup solutions are interesting they're outside the reality for individuals and small businesses. If we were to poll the users in this forum I would be surprised if the majority of them were using something other than a RAID configuration with some form of backups...just as the poster you're referring to was doing.

The frequency of his backups may not have been sufficient but then backups, as we can see as very important, tend to take a backseat to getting real work done (i.e. as in the work that pays the bills).

I think a more appropriate question to ask of the forum is what are individuals / small businesses doing for their backup and disaster recovery.
 
Rebuilding any kind of raid except for perhaps mirror is quite an issue that many RAID fanboys conveniently don't mention. Most enterprise data centers I know of simply rotate cheap drives in an out on a routine basis, so they rarely have to deal with an unexpected failed drive.

It seems that, when one objectively runs the numbers and failure scenarios, for the pro that needs less than say 10TB of project storage, RAIDs are just not useful at all. RAID can be a feel good thing however, but today its usefulness is subjective.... a mirror set up is most useful for the convenience of rotating drives through to off site.. but the more enterpri$$$e like you get, there are better solutions.
 
Firstly, if you rely on your data for your income, then you must follow the holy trinity of:
  1. Daily reliable backups.
  2. Off-site replication.
  3. Disaster recovery.
If you plan these 3 items then you can be sure of your data integrity.

The hard part is what are your requirements. If you backup solely to the cloud, you essentially cover all your bases and have a fairly low cost.

But when (not if) you do have a failure, how quickly do you need the data recovered? Cloud solutions usually mean a very slow restoration process which depending on the size can take days to replicate back into your environment. Can you and your projects afford to sit and wait for several days as that happens?

A multi-tiered approach is often the best strategy, cost is obviously a major factor, but luckily due to the proliferance of consumer grade NAS devices from Enterprise grade manufacturers we now have options in all price points so we can find a solution for everyone.

A decent workstation environment should have at minimum a daily replication period to local storage (local being within the same building).

The older Mac Pro's can hold 4 Hdds, so that is going to be the quickest place for you to perform a restoration from to think about keeping a complete replica there ready to take over incase of a drive failure you can use RAID in your boot and data drives for this. Recent advances in solid state storage improved speed and reliability so this is no longer so important but we still need to make sure we have local backups and remember the mantra (RAID is not backup).

All your local files and data must be backed up to a separate storage device.

What this is depends on the number of computers, and how much money you want to spend.

QNAP had a nice range of models
so you can backup multiple computers over the local network many models (even lower end) support 10Gb so you can have very good speeds (I am not sure if anyone has gotten 10Gb cards working in Mac Pros though).

Once the data is saved to the backup device, you can now think about off-site replication, if you have another office, you can replicate data to a device in that office and vice-versa. Or you can replicate to a cloud storage provider. Either way, what you are protecting against if physical destruction of the data at a single location (fire, earthquake, meteor).

If you replicate to another office, the data is still "local" to you so you can get the data back into a working environment quickly.
[doublepost=1477233075][/doublepost]
Aren't you being overly hard on the guy? It sounds as if he was doing what he could within the constraints he had. He had a RAID5 configuration and was performing some level of backups. He just had a bit of bad luck in that he had multiple drive failures at the same time in his RAID5 configuration.

While enterprise backup solutions are interesting they're outside the reality for individuals and small businesses. If we were to poll the users in this forum I would be surprised if the majority of them were using something other than a RAID configuration with some form of backups...just as the poster you're referring to was doing.

The frequency of his backups may not have been sufficient but then backups, as we can see as very important, tend to take a backseat to getting real work done (i.e. as in the work that pays the bills).

No! This is exactly why everyone needs to actively think and prepare for this.

Budget is not a reason for a bad strategy, it is an excuse and will always bite you when you are unprepared.

A good backup strategy is not just for enterprise. You can backup, and replicate your data on a budget. The trade off to cost is speed of recovery.

The bigger question is can you afford to not have a good data backup and recovery solution in place.

I build and manage virtual machine clusters around the world, the hardware and solutions are different but the concepts are the same and can be done at nearly any price point.
 
No! This is exactly why everyone needs to actively think and prepare for this.

Budget is not a reason for a bad strategy, it is an excuse and will always bite you when you are unprepared.

A good backup strategy is not just for enterprise. You can backup, and replicate your data on a budget. The trade off to cost is speed of recovery.

The bigger question is can you afford to not have a good data backup and recovery solution in place.

I build and manage virtual machine clusters around the world, the hardware and solutions are different but the concepts are the same and can be done at nearly any price point.
Easy to say, difficult to implement. The reality is the individual / small business owner is unlikely to have the expertise and/or money to implement a solution which would prevent any data loss.

Most NAS's are little more than a network attached RAID. It's subject to the same types of failures as direct attached storage. The same failure which caused the individual of whom you've based this thread to lose data.

The issue was lack of regular backups. We don't know his work environment, we don't know his configuration, and we don't know his data backup strategy. What we do know is he took what I consider to be reasonable precautions based on his level of knowledge and / or cost (I don't know him but I get the idea he's either an individual or working in a very small organization without dedicated IT). He protected his workstation with RAID5 (reasonable) and had some level of backup (reasonable). The only questions are: What was the backup strategy and how could he have improved on it?

I personally use CrashPlan. I have it installed on my main PC and Macs and they're configured to backup to the Mac Pro I use as a virtualization server (as it is on 24/7). My family and friends are configured to backup over the Internet to this same system (providing them nightly backups which have the benefit of being offsite). I have no such target so this backup strategy provides only system level protection. In order to work around that I regularly, when I deem it appropriate, make a backup to an external drive and take it off site. This means I am at risk of losing data should my house burn to the ground (or some other similar event). But I've weighed the risks and the majority of what I would want to keep is stored offsite even if not current.

This strategy also doesn't protect me against ransomware as there is no versioning. If I don't catch it the encrypted files overwrite my plaintext files rendering my backups worthless (though I do have the offsite which has most of the important stuff backed up and I, on a monthly basis, archive the CrashPlan directories of my family and friends in the event they get hit so they won't lose everything).

I could expand upon this strategy because it doesn't provide me the ability to recover everything in the event of an event that wipes out my house. But I've weighed the risks versus cost versus management and I've accepted the risk associated with this strategy.
 
Easy to say, difficult to implement. The reality is the individual / small business owner is unlikely to have the expertise and/or money to implement a solution which would prevent any data loss.

Most NAS's are little more than a network attached RAID. It's subject to the same types of failures as direct attached storage. The same failure which caused the individual of whom you've based this thread to lose data.

The issue was lack of regular backups. We don't know his work environment, we don't know his configuration, and we don't know his data backup strategy. What we do know is he took what I consider to be reasonable precautions based on his level of knowledge and / or cost (I don't know him but I get the idea he's either an individual or working in a very small organization without dedicated IT). He protected his workstation with RAID5 (reasonable) and had some level of backup (reasonable). The only questions are: What was the backup strategy and how could he have improved on it?

I personally use CrashPlan. I have it installed on my main PC and Macs and they're configured to backup to the Mac Pro I use as a virtualization server (as it is on 24/7). My family and friends are configured to backup over the Internet to this same system (providing them nightly backups which have the benefit of being offsite). I have no such target so this backup strategy provides only system level protection. In order to work around that I regularly, when I deem it appropriate, make a backup to an external drive and take it off site. This means I am at risk of losing data should my house burn to the ground (or some other similar event). But I've weighed the risks and the majority of what I would want to keep is stored offsite even if not current.

This strategy also doesn't protect me against ransomware as there is no versioning. If I don't catch it the encrypted files overwrite my plaintext files rendering my backups worthless (though I do have the offsite which has most of the important stuff backed up and I, on a monthly basis, archive the CrashPlan directories of my family and friends in the event they get hit so they won't lose everything).

I could expand upon this strategy because it doesn't provide me the ability to recover everything in the event of an event that wipes out my house. But I've weighed the risks versus cost versus management and I've accepted the risk associated with this strategy.

Thanks ITguy2016. You have hit the nail directly on the head, but I just want to stress it's importance.

Daily backups are vital and should not be ignored.

The lack of knowledge and/or expertise issue will hopefully be addressed by people posting their solutions here (the purpose of this thread).

I also appreciate that every office is different and every solution is different, but the core 3 elements must be thought about and addressed.

There is no magic to an enterprise level SAN unit beyond redundant controllers and fabric. There are plenty of cheap backup units that will suffice, it is important that the backup unit doesn't become the primary location for saving older files. Backup means a copy of the data in another device. I have seen people say my backup drive died and be screwed because they had been not backing up but just storing the data to an external drive and referring to it as a backup.

Apple's own time machine is a consumer grade backup software, it or something like superduper can be the first step in the backup chain.

No one should ever say I don't have automated daily backups because of expense or lack of knowledge. Yes, this is an unsexy topic, but vital if you are running a business.
 
I have found a combo of Time Machine and CarbonCopy clone to be nicely redundant. CCC is much faster at restoring, while TimeMachine is always more current. Both can have problems.

I have restored from both more than once.
 
  • Like
Reactions: jbarley
My backup workflow consist of;
Raid 0 4tb + OS Drive + 2x 3tb --> Raid 5 10tb --> Time Machine Seagate 8tb Drive (gets swapped every year and old one gets kept for backup)

then i use - https://www.backblaze.com - So cheap its 50 a year
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.