Backup A. B. C.
You need to have three copies of your files:
A – this is your working copy.
B – this is your on-site backup, which protects you from mistakes and failures within your working copy. It is quick to access.
C – this is your off-site backup, which is slower, but protects you from burglary, fire, flood, war and other disasters.
And that’s it.
You can find some concrete solutions described further below.
Some important details
All copies should be encrypted. You don’t want burglars or backup providers to snoop through your files. If A-copy is in your PC or laptop, it should be encrypted too. At least with BitLocker or similar. VeraCrypt is worth considering. For off-site copies, they should be client-side encrypted (“zero knowledge backup” – backup provider should not be able to read your files).
You must be able to recover your cloud access password and encryption key even if your house is burned down to the ground. Consider some password vaults, like KeePass or KeePass XC and have their databases replicated across multiple devices (laptop, mobile phone, cloud backup).
The C-copy should be as off-site as possible, even on another continent. It should also be in a place with decent law system, human rights record and privacy regulations (you don’t want your files seized, because one of your pictures is not in line with some local regulations or political agenda – encryption helps here too).
The C-copy should be on a platform offered by major backup company, with decent customer support. Look past the list price and be wary of unexpected costs, such as cost of restoring data. Some companies offer to send you your data on a hard drive.
One of the backup tasks (A-copy to B-copy or B-copy to C-copy) should not be performed instantly. If you delete something by mistake you don’t wont this mistake to be instantly replicated to your backup. Some cloud providers keep files you delete in a “trash” for some time. This could be useful.
No backup is 100% reliable. Not even cloud backup (look into their specs, Backblaze for example declares 99.999999999% “annual availability”). You should have B-copy even if you have a cloud backup as a C-copy. And without B-copy, if you delay backup as suggested above, there will be no backup of your latest changes for some time.
Test your backup. Never trust backup until you verified you can retrieve data. Do this periodically.
I have a Synology / QNap / WD / other NAS, isn’t that enough?
NAS alone is not a backup, see below.
What is a NAS good for?
NAS is not a backup, if it is the only place your data is. If you drop the NAS from the table, quite likely more drives will fail than the NAS can survive and you will lose your data. If it will get accidentally erased or stolen, you are left with nothing. So once again: NAS alone is not a backup.
What do NAS really offer:
- Ability to survive failure of one (or two) drives without damage to your data, and even without interrupting your work – in certain setups. But if more drives fail – data is gone.
- Ability to hot-swap failed drive with new one without interrupting your work.
- Ability to create single, large file system over multiple drives. Need 50TB of storage? Buy 4 12TB drives and let NAS create single file system over them.
- Ability to expand storage without interrupting work.
- Faster speed in certain setups – with stripping and fast connection between NAS and PC or laptop you may achieve faster transfers than with single drive.
- Software for automation of backups, often with support for using cloud backups as C-copy.
So what are NAS good for? They are good solution if you need more space than can be provided by single drive for A-copy or B-copy.
They can also save you from an interruption of your work when one or two drives will fail (in certain setups). And all drives eventually fail (see below).
NOTE: Most popular NAS solutions relay on something called RAID for it’s hard drive management. You can get most of above benefits running normal computer with multiple drives and RAID software or hardware. E.g. Linux and it’s software RAID are very popular for NAS solutions – both custom and commercial, like Synology (DSxxx series uses Linux software RAID).
All drives fail, be ready for that
Whatever solution you have, be ready for a drive failure. Drives fail. Backblaze reports that 1 percent of their drives fails in a single quarter.
Some drives fail quickly within warranty, others after years. It’s random. Accept this and decide what you will do when one of your drives will fail. Even if your NAS is configured to survive drive failure, you don’t really want to run your NAS without spare drive. Also, rebuilding RAID after replacing drive takes time – until it is done, you have no resiliency. So it is best to have a spare drive handy.
Solution 1: if your data fits one drive
Your data fits a single drive?
A-copy – HDD or SSD in your PC or laptop.
B-copy – portable USB drive.
C-copy – cloud backup such as Google Drive, One Drive, iDrive, Backblaze and many others.
Some software for copying from A-copy to B-copy would be handy, but you can use Total Commander as well. Just do this on regular basis. Also, some cloud solutions support local backups too (e.g. iDrive).
Solution 2: Setup if you need more capacity than one drive can offer
Let’s say you are a pro photographer and have 30TB of media collected during your career.
A-copy – NAS with 4 12TB drives with RAID-5. This offers one-drive failure resiliency.
B-copy – NAS with 4 12TB drives with RAID-5. This offers one-drive failure resiliency.
C-copy – cloud backup.
Solution 2b: Cheaper middle-ground with one NAS (no B-copy)
You may consider skipping B-copy, but if you are very unlucky to lose more than one drive in NAS and hit that 0.000000001% failure rate of your cloud storage at the same time, you will lose everything. You need to find your own balance between cost and reliability.
Remember also that B-copy is to allow you to delay replication of changes to C-copy. Because on the one hand you want to have your latest changes backed up as soon as possible, but on the other hand, you want your latest changes delayed in case of an unrecoverable mistake (such as deleting files). Some backup solutions offer limited ability to recover deleted / overridden files.
Let’s say again you are a pro photographer and have 30TB of media collected during your career.
A-copy – NAS with 4 12TB drives with RAID-5. This offers one-drive failure resiliency. (Or NAS with 5 12TB drives with RAID-6. This offers two-drive failure resiliency.)
B-copy – NONE.
C-copy – cloud backup with “trash” support – ability to restore deleted files.
Benefits are obvious – one NAS less, 4 drives instead of 8. Much cheaper solution.
What are downsides? If you make a mistake and need to recover data from your backup, you need to take it from the cloud. This will be slower and and less convenient that from local backup. Also, depending on your settings, cloud might not be updated instantly, so you may be only able to recover data from yesterday.
What do I use – at the desk?
I have around of 7TB of personal data. Large part of it is my photo/video collection.
All my files sit on a custom-built NAS, with 3x4TB HDD in RAID-5 configuration (single drive failure resiliency). This is A-copy.
Data from the NAS is backed up nightly to iDrive. This is C-copy.
No B-copy. And before you’ll ask: yes, I already had to restore some files from iDrive. Not as fast as from local backup, but it works.
I work on a desktop PC. Usually I use files directly from the NAS, but occasionally I copy current work to PC (keeping it on NAS).
Lightroom Collection is a pain point, as it can’t be used from the NAS. I have a copy on my PC. It is backed up to the NAS periodically.
If I will need to expand my storage in the future, I will have these options:
- Add another 4TB drive and get 12TB of storage with 1-drive failure resiliency.
- Add two 4TB drives and get 12TB of storage with 2-drive failure resiliency (this is RAID-5 to RAID-6 conversion which could be very time consuming).
- Buy three 8TB drives, build new RAID-5 and get 16TB with 1-drive failure resiliency (this would require restoring data from old 4TB drives or from the cloud backup).
What do I use – on the go?
When I’m away, I’m working on a laptop PC. In the evenings I’m backing up my photos/videos to the laptop, keeping them on the source SD/microSD cards – data is always in two places.
Also, I’m backing the laptop up to iDrive. Here my experience is that hotel WiFi is often not fast enough to perform cloud backup overnight if I have video files.
If I need more space than my laptop can offer, I’m using portable SSD drive.
When I’m back home, I’m copying everything to the NAS and clean the laptop and SD/microSD cards only after cloud backup is complete.
Conclusion
Having a backup solution in this era of information is a must. With tools available it is pretty simple to chose one and keep our data safe. And this is one of these things we shouldn’t be postponing. Data recovery is costly and peace of mind is invaluable.