How do you backup?

I'm more of a hobbyist that does occasional pro work ad hoc. I don't really trust consumer grade RAID units anymore. I have had two different 2-drive external HDD units fail over the past few years. I had both set up as RAID 1.

When a drive went bad in the first unit, I found that I could not just pull out the remaining drive, put it in a reader, and get files off it using Windows. I had to buy software to read the file structure and retrieve the files, but at least I got my data back.

When a drive went bad in the second unit, I replaced the dead drive with a replacement unit from the manufacturer which was supposed to initiate an automatic reconstruction... but that didn't happen and it just sat there giving me an error. I ended up pulling out the remaining drive and was able to read it in Windows directly, but then that drive also started to respond slowly and I ended up only retrieving about 10% of my data before that drive died.

Now I have 4 independent 4TB drives in my desktop (D, E, F, G drives), which I use in two 2-drive pairs. I only directly work with files on my D and F drives. Then every evening I have Cobian Backup copy all the changes from D to E, and F to G. This doesn't protect so well against data corruption, but Cobian doesn't actually mirror, so if a file on one drive gets corrupted, it stands to reason that the matching file on the second drive might be okay.

I also use Backblaze for ongoing cloud backup.

I do have an older 12TB RAID that I bought about 6 years ago, set up as RAID 5 and just put copies of my files on that one since I have it, but I don't really consider it as part of my main backup routine anymore.
 
I'm more of a hobbyist that does occasional pro work ad hoc. I don't really trust consumer grade RAID units anymore. I have had two different 2-drive external HDD units fail over the past few years. I had both set up as RAID 1.
That is why I went for an enterprise grade RAID controller. Not only it is a lot flexible and secure, even if the RAID controller dies I could just put in another controller of the same brand (Not even the same model), and it will read the configuration and have it up in a second. Also using enterprise grade hard drive helps a lot.
When a drive went bad in the first unit, I found that I could not just pull out the remaining drive, put it in a reader, and get files off it using Windows. I had to buy software to read the file structure and retrieve the files, but at least I got my data back.

When a drive went bad in the second unit, I replaced the dead drive with a replacement unit from the manufacturer which was supposed to initiate an automatic reconstruction... but that didn't happen and it just sat there giving me an error. I ended up pulling out the remaining drive and was able to read it in Windows directly, but then that drive also started to respond slowly and I ended up only retrieving about 10% of my data before that drive died.
Ohh that is sad.
Now I have 4 independent 4TB drives in my desktop (D, E, F, G drives), which I use in two 2-drive pairs. I only directly work with files on my D and F drives. Then every evening I have Cobian Backup copy all the changes from D to E, and F to G. This doesn't protect so well against data corruption, but Cobian doesn't actually mirror, so if a file on one drive gets corrupted, it stands to reason that the matching file on the second drive might be okay.

I also use Backblaze for ongoing cloud backup.

I do have an older 12TB RAID that I bought about 6 years ago, set up as RAID 5 and just put copies of my files on that one since I have it, but I don't really consider it as part of my main backup routine anymore.
 
I'm more of a hobbyist that does occasional pro work ad hoc. I don't really trust consumer grade RAID units anymore. I have had two different 2-drive external HDD units fail over the past few years. I had both set up as RAID 1.
That is why I went for an enterprise grade RAID controller. Not only it is a lot flexible and secure, even if the RAID controller dies I could just put in another controller of the same brand (Not even the same model), and it will read the configuration and have it up in a second. Also using enterprise grade hard drive helps a lot.
Good point. I just checked prices on LSI Logic RAID controllers and they are a lot less money than I was expecting.

Thank you. This is definitely something I will consider when I want to upgrade.

For hard drives I totally agree on getting server or enterprise grade drives.
 
Last edited:
I do it "caveman style". Shoot on dual 128g sd cards, download to the laptop then copy that folder to twin 5g WD portable hard drives, when the 1tb hard drive on the laptop starts to fill up in a month or so I'll delete those files. So far I've had one hard drive failure probably due to the high humidity conditions at the storage box place-no longer used. Someday I hope to go to SSD's for backup.

I don't do the cloud too much (a lot of cloud thieves) and Raid/NAS seems expensive and complicated, I'm just an amateur after all... Other than the bother of storing multiple hard drives (several from 500g to 5tb) it's been pretty easy and redundant. Someday SSD will become more affordable and I'll migrate to them then.
 
My 8TB Samsung QVO internal Sata SSD just arrived today and I will mount it in my system this Sunday.

So all of my files will be on that 8TB SSD with plenty of room for future years. That is 4 times faster than my spinning 8TB HDD.

I hope that in two more years we will have 8TB M.2 NVMe PCIe SSDs for 500 bucks. That 8TB SATA drive was 900. The M.2 SSD drive is 7 times faster than the SATA SSD. But the SATA SSD is 4 times faster than a spinning HD, so I'm happy.

I also bought a 2TB Seagate FireCuda 510 M.2 PCIe NVMe SSD for my OS and programs.

The challenge will be to switch from my old Boot 1 TB SSD and my spinning hard drive to these two new SSDs.

I will have to reload windows and all my programs on the new 2TB boot drive and then transfer all my 5 TB of image files to the new 8TB SSD. I have to make sure it all works and that my LR catalog links up to that new disk from that new boot drive.

Shouldn't be a problem.

But it might be.....
 
I can recommend the M-Disk. It's rated to 1000 years and nothing else comes close. It's really durable and 25-100 Gb discs are practical for usage. Prices are also good.
 
This is my full image management routine from start to finish:
  • Image Management 1 - Preparing the storage media for travelling and ensuring reliable and fast operation of storage media and the camera
  • Image Management 2 - Culling, editing, and backing up the images while on the road
  • Image Management 3 - Safeguarding your images once you get them home
I want to especially draw your attention to the "Easy Restoration" section in Image Management 3; it's a step missed by many people, and yet its impact can be devastating at the worst possible time. Here is that excerpt:

Easy Restoration

The restoration process is very important to getting your data back. The only way to really know if your end-to-end redundancy, backup and restore process works is to test it. You must test your restoration process to be confident you’ll be able to get your data back. To test the end-to-end process, I recommend doing the following:
  1. Load the same set of files onto your NAS into four directories, named:
    • 1-day
    • 7-days
    • 30-days
    • 90-days
  2. Allow your backup processes to complete, onsite and offsite
  3. After one day, delete all four folders
  4. Test your restore capability by attempting to:
    • Restore the 1-day folder right away
    • Restore the 7-days folder after a week
    • Restore the 30-days folder after a month
    • Restore the 90-days folder after a three months
Doing the restore test in this way will catch the subtleties such as the impact some settings like versioning have on your restore capabilities. For example, when I accidentally deleted the folders and files with the out-of-control USB import, had I not caught it right away, without versioning turned on in the cloud solution, the NAS/B2 synchronization would have eventually deleted those files from the cloud backup too, leaving only the older offsite backup as a viable solution.

--
How to rescue an image
Capture One Pro quick workflow
Travel and photography blog: fcracer.com
 
Last edited:
I can recommend the M-Disk. It's rated to 1000 years and nothing else comes close. It's really durable and 25-100 Gb discs are practical for usage. Prices are also good.
If I use M-Disk, I'll end up needing 180 x 100GB M-Disks. Now imagine the loading / unloading time plus actual backup, and organizing data in 100GB blocks so I can back them up?

Life is short.
 
This is my full image management routine from start to finish:
  • Image Management 1 - Preparing the storage media for travelling and ensuring reliable and fast operation of storage media and the camera
  • Image Management 2 - Culling, editing, and backing up the images while on the road
  • Image Management 3 - Safeguarding your images once you get them home
I want to especially draw your attention to the "Easy Restoration" section in Image Management 3; it's a step missed by many people, and yet its impact can be devastating at the worst possible time. Here is that excerpt:

Easy Restoration

The restoration process is very important to getting your data back. The only way to really know if your end-to-end redundancy, backup and restore process works is to test it. You must test your restoration process to be confident you’ll be able to get your data back. To test the end-to-end process, I recommend doing the following:
  1. Load the same set of files onto your NAS into four directories, named:
    • 1-day
    • 7-days
    • 30-days
    • 90-days
  2. Allow your backup processes to complete, onsite and offsite
  3. After one day, delete all four folders
  4. Test your restore capability by attempting to:
    • Restore the 1-day folder right away
    • Restore the 7-days folder after a week
    • Restore the 30-days folder after a month
    • Restore the 90-days folder after a three months
Doing the restore test in this way will catch the subtleties such as the impact some settings like versioning have on your restore capabilities. For example, when I accidentally deleted the folders and files with the out-of-control USB import, had I not caught it right away, without versioning turned on in the cloud solution, the NAS/B2 synchronization would have eventually deleted those files from the cloud backup too, leaving only the older offsite backup as a viable solution.
True, testing out the backups so you know it works in the event of a disaster is crucial. I backup about 18TB of data, so it is not possible for me to have all the versioning copies in one location. I do them in different drives / RAID arrays. Every other month, I did a whole surface check of all the drives (read and write) to ensure there are no media errors such as bad blocks and such. This surface test is a three phase process. It reads the data from every sector, then writes it back to the sector and then reads it back again to ensure sector health and data integrity.

The main RAID array goes through monthly routine check of drive surface, and a consistency check to ensure datas are not corrupted anywhere. If it is, it reconstructs the data from parity, and logs it.

A backup is no good if it is not retrieveable, and maintaining good backup health is just as important.

I am going through full surface checks of all the drives I have, and I still have 16 x 10TB, 12 x 8TB, 2 x 16TB and 1 x 14TB drives to go through 😔 such a long and tedious process.
 
nananana-boo-boo, I'm on an 8TB SSD and synced to two spinning 8TB HDDs now (as back up), so you guys can take your RAID and stick it.....

Just kidding..... 😁

And yes .... I have a copy of the sync off site.....

Now I'm dreaming of the day that the 8TB SSD is an M.2 for 500 bucks and on the Motherboard (instead of an SSD sata drive that costs 900 bucks)!

Now that will be sweeeeeeeeet.

Raid, Raid, the witch is dead, the wicked witch oh witch is dead, is dead is dead....

(Wizard of Oz tune...).

You guys need to sing along with me....
 
nananana-boo-boo, I'm on an 8TB SSD and synced to two spinning 8TB HDDs now (as back up), so you guys can take your RAID and stick it.....

Just kidding..... 😁

And yes .... I have a copy of the sync off site.....

Now I'm dreaming of the day that the 8TB SSD is an M.2 for 500 bucks and on the Motherboard (instead of an SSD sata drive that costs 900 bucks)!

Now that will be sweeeeeeeeet.

Raid, Raid, the witch is dead, the wicked witch oh witch is dead, is dead is dead....

(Wizard of Oz tune...).

You guys need to sing along with me....
Hahahaha

Greg, dead or alive, RAID or normal HDD or SSD, everything needs maintenance. That is what we were talking about. Once or twice a month, you should do full surface checks on your drives (If possible both read / write test), to make sure the drives do not have any weak sectors or blocks. This goes for SSD / HDD both.

A corrupted / incomplete backup due to drive fault is, well, not a good backup.
 
nananana-boo-boo, I'm on an 8TB SSD and synced to two spinning 8TB HDDs now (as back up), so you guys can take your RAID and stick it.....

Just kidding..... 😁

And yes .... I have a copy of the sync off site.....

Now I'm dreaming of the day that the 8TB SSD is an M.2 for 500 bucks and on the Motherboard (instead of an SSD sata drive that costs 900 bucks)!

Now that will be sweeeeeeeeet.

Raid, Raid, the witch is dead, the wicked witch oh witch is dead, is dead is dead....

(Wizard of Oz tune...).

You guys need to sing along with me....
Hahahaha

Greg, dead or alive, RAID or normal HDD or SSD, everything needs maintenance. That is what we were talking about. Once or twice a month, you should do full surface checks on your drives (If possible both read / write test), to make sure the drives do not have any weak sectors or blocks. This goes for SSD / HDD both.

A corrupted / incomplete backup due to drive fault is, well, not a good backup.
Surface checks on SSD may reduce the life of SSD, AFAIK. Maybe a SMART test is better?

A good backup strategy preserves all old copies of a file. If an image gets corrupted, you should be able to recover it from backup, unless the backup disk gets corrupted :(.

Multiple backups, stored all over the world, sounds safest. I guess that is done with cloud storage.
 
nananana-boo-boo, I'm on an 8TB SSD and synced to two spinning 8TB HDDs now (as back up), so you guys can take your RAID and stick it.....

Just kidding..... 😁

And yes .... I have a copy of the sync off site.....

Now I'm dreaming of the day that the 8TB SSD is an M.2 for 500 bucks and on the Motherboard (instead of an SSD sata drive that costs 900 bucks)!

Now that will be sweeeeeeeeet.

Raid, Raid, the witch is dead, the wicked witch oh witch is dead, is dead is dead....

(Wizard of Oz tune...).

You guys need to sing along with me....
Hahahaha

Greg, dead or alive, RAID or normal HDD or SSD, everything needs maintenance. That is what we were talking about. Once or twice a month, you should do full surface checks on your drives (If possible both read / write test), to make sure the drives do not have any weak sectors or blocks. This goes for SSD / HDD both.

A corrupted / incomplete backup due to drive fault is, well, not a good backup.
Surface checks on SSD may reduce the life of SSD, AFAIK. Maybe a SMART test is better?

A good backup strategy preserves all old copies of a file. If an image gets corrupted, you should be able to recover it from backup, unless the backup disk gets corrupted :(.

Multiple backups, stored all over the world, sounds safest. I guess that is done with cloud storage.
Write test will reduct the SSD write limit, but most modern drives have such high write endurance, that a complete drive write once or twice a month is practically nothing. Read test won't affect the SSD at all.

For example, the SSD Greg has, has a write endurance of 2880 TB, or 0.3 drive write per day. That is huge, and 8TB from that every month or two is basically nothing.

Yes, that is we were talking about, checking backup drive integrity in the case a restore is needed :)

Yeah cloud can really help do that, only downside being the slow retrieval unless someone has those Google 1GBit lines haha
 
I’ve been thinking about tape drives lately. Feeling nostalgic. Does anyone know the current state of the art archival tape tech? Has the price come down or still at enterprise expense levels? Would be awesome to archive TB of data to a tape to store offsite instead of dealing with clunky HDD (which top out at ~18TB).
 
I’ve been thinking about tape drives lately. Feeling nostalgic. Does anyone know the current state of the art archival tape tech? Has the price come down or still at enterprise expense levels? Would be awesome to archive TB of data to a tape to store offsite instead of dealing with clunky HDD (which top out at ~18TB).
Top out at 18 TB? Tape drives? How much data do you have? Wow....
 
I’ve been thinking about tape drives lately. Feeling nostalgic. Does anyone know the current state of the art archival tape tech? Has the price come down or still at enterprise expense levels? Would be awesome to archive TB of data to a tape to store offsite instead of dealing with clunky HDD (which top out at ~18TB).
Top out at 18 TB? Tape drives? How much data do you have? Wow....
 
I use Storage Craft’s Image Protect

A serious backup and disaster recovery tool. Even did a bare metal restore as a test Extremely reliable & flexible solution.
 
I’ve been thinking about tape drives lately. Feeling nostalgic. Does anyone know the current state of the art archival tape tech? Has the price come down or still at enterprise expense levels? Would be awesome to archive TB of data to a tape to store offsite instead of dealing with clunky HDD (which top out at ~18TB).
I used to use LTO, but the standard kept changing and drivers for older drives stopped being supported in newer operating systems.
 
nananana-boo-boo, I'm on an 8TB SSD and synced to two spinning 8TB HDDs now (as back up), so you guys can take your RAID and stick it.....

Just kidding..... 😁

And yes .... I have a copy of the sync off site.....

Now I'm dreaming of the day that the 8TB SSD is an M.2 for 500 bucks and on the Motherboard (instead of an SSD sata drive that costs 900 bucks)!

Now that will be sweeeeeeeeet.

Raid, Raid, the witch is dead, the wicked witch oh witch is dead, is dead is dead....

(Wizard of Oz tune...).

You guys need to sing along with me....
Hahahaha

Greg, dead or alive, RAID or normal HDD or SSD, everything needs maintenance. That is what we were talking about. Once or twice a month, you should do full surface checks on your drives (If possible both read / write test), to make sure the drives do not have any weak sectors or blocks. This goes for SSD / HDD both.

A corrupted / incomplete backup due to drive fault is, well, not a good backup.
Surface checks on SSD may reduce the life of SSD, AFAIK. Maybe a SMART test is better?

A good backup strategy preserves all old copies of a file. If an image gets corrupted, you should be able to recover it from backup, unless the backup disk gets corrupted :(.

Multiple backups, stored all over the world, sounds safest. I guess that is done with cloud storage.
Write test will reduct the SSD write limit, but most modern drives have such high write endurance, that a complete drive write once or twice a month is practically nothing. Read test won't affect the SSD at all.

For example, the SSD Greg has, has a write endurance of 2880 TB, or 0.3 drive write per day. That is huge, and 8TB from that every month or two is basically nothing.

Yes, that is we were talking about, checking backup drive integrity in the case a restore is needed :)

Yeah cloud can really help do that, only downside being the slow retrieval unless someone has those Google 1GBit lines haha
Many cloud services will ship recovery data via UPS. Mine ships arrays.
 
Jim I did a straight windows copy of my D drive (which was my old primary main spinning 8TB HDD) to my new G Drive - 8TB Samsung QVO SSD, and then today I used GoodSync to set up a job which now goes from G to D, meaning now my old primary drive that I was syncing from is now the recipient of the sync from the drive I copied to (G).

I went in LR and edited one raw file. Then I synced from G to D (D being my old primary HDD that is now a backup drive being synced to).

GoodSync immediately recognized all the files that were on D and synced that one change to an XMP file (sidecar) from that one edit off the G Drive SSD.

I guess that mean no files were corrupted on the copy from my D to my F. If they were, I'm screwed because now I am operating off the new 8TB SSD as my main single drive and syncing to 5 backup drives.

I didn't do that checksum you suggested. Couldn't figure it out. It scared me.

--
Greg Johnson, San Antonio, Texas
https://www.flickr.com/photos/139148982@N02/albums
 
Last edited:

Keyboard shortcuts

Back
Top