Do you back-up your images or copy them?

Status
Not open for further replies.

MMarz

Active Member
Joined
Nov 13, 2007
Messages
469
I have been making copies of my Master images, thinking it would be easier to restore an image or image set if I ever needed to, but truth told, I have never needed to. Since I have and use Acronis True Image for other purposes I am contemplating creating backups instead of copies..

I am curious what the rest of you do?
 
I'm not sure how you are using Acronis True Image, but I think one of the ways is an uncompressed backup copy. I have three backup processes that are active on my local image files. [list type=decimal]
[li]Carbonite: A Cloud backup service. Anytime a master image changes, Carbonite makes a new copy.(Mine are DNGs which LR manages the XMP metadata). [/li]
[li]AllWay Sync. Does incremental copies of the local drive image files on MWF.[/li]
[li]RAID1: My backup copies and my non-local image files are stored on a NAS. this NAS has a RAID1 Mirror that is continuously in sync. [/li]
[/list] Every one of these backup processes produces an uncompressed copy which I can access any time I need to.

The reall question is not WHETHER you will need to use your backups to restore, but When.

In the last two years, my wife has had two HDs fail and we have had one NAS backup HD fail
 
I manually copy my files to two other external hard drives( total of three copies). I know this is a "caveman version of RAID" but it works.

Bob- the slide shooter at heart
 
Bob, you may think of it as a "caveman approach" to disk mirroring, but it has one major advantage over traditional RAID 1 -- corruption on one side of the mirror won't be automatically propagated to the other side.

I do the same thing at import time with my original files -- I don't want to use RAID 1 for this very reason. I do run a job regularly that compares each of the three copies to detect corruption.
 
I failed to expand on my current backup workflow..

My Masters are ingested to the an internal data drive (converted to DNG's) that holds the working copies of the two most recent years. An initial cull is done and immediate rejects are marked and deleted. The remaining pics are copied to an external drive which holds all images from all years. Cards can then be formatted and edits can begin.

After each edit session I make sure to write the xmp data to the DNG's and one a week or so, I use Beyond Compare to audit the copies and update those that have changes done to them.

Backblaze then does another backup to the clouds of all images on the fly. I have another external that periodically gets update of all images manually mirrored and is stored off-site.

My question about backup versus copies it targeted at replacing the off-site external with a compressed backup instead of copies. Acronis runs a backup of all the basic stuff I need backed up, including the LR catalog backup. Unrelated, my transient files are all sync'd using Dropbox.
 
[quote author=Mark Sirota link=topic=975'.msg66'16#msg66'16 date=1273'27961]
Bob, you may think of it as a "caveman approach" to disk mirroring, but it has one major advantage over traditional RAID 1 -- corruption on one side of the mirror won't be automatically propagated to the other side.

I do the same thing at import time with my original files -- I don't want to use RAID 1 for this very reason. I do run a job regularly that compares each of the three copies to detect corruption.
[/quote]

Mark,

What program are you using to detect data corruption?

--Ken
 
I would like to know how to script this on my MacPro tough I do not know how to use Automator or AppleScript yet...

For Windows, the only way I know to verify is by using the -v switch for the copy or xcopy commands.
 
That shellscript can do for me. I don't care about the way to do it be it in terminal or otherwise.
 
You people are obviously computer literate but, I'm a bit confused as I couldn't tell a shell script from a pipe wrench. But, checking for corrupted photo files sounds like common sense. Is there a simple way that people with minimal computer knowledge can accomplish this? Thanks.
 
I would think that modern fancy backup software could do this, but I don't know for sure.

This method works for me specifically because I have three identical write-once folder trees which include all my originals, and I don't write XMP data out to sidecars (or DNGs) So these folder trees never change, and can be compared periodically. If your workflow doesn't work that way, my method won't work for you.

Denis, if it'll work for you, I can share the relevant part of my script. The meat of it is:
Code:
diff -rq /Volumes/Photos-1/Pictures/Originals /Volumes/Photos-2/Pictures/Originals
It sends any the output to me by e-mail. I've been meaning to have it do something fancier with the output -- there isn't any unless there's a problem, so I don't want to risk losing it in my e-mail.
 
[quote author=Mark Sirota link=topic=975'.msg66'16#msg66'16 date=1273'27961]
Bob, you may think of it as a "caveman approach" to disk mirroring, but it has one major advantage over traditional RAID 1 -- corruption on one side of the mirror won't be automatically propagated to the other side.[/quote]

Great! When someone makes fun of me, I can use this as a counter argument.

Bob- the slide shooter at heart
 
[quote author=Mark Sirota link=topic=975'.msg66'16#msg66'16 date=1273'27961]
Bob, you may think of it as a "caveman approach" to disk mirroring, but it has one major advantage over traditional RAID 1 -- corruption on one side of the mirror won't be automatically propagated to the other side.
[/quote]

This is not actually the case.... If an OS system call writes corrupted data to the mirror it will effect both sides. This is also the most common corruption in my experience. If a disk suffers from a problem/failure and causes corruption then the other disk in the mirror will be ok.
 
[quote author=Digital Reef link=topic=975'.msg66'66#msg66'66 date=1273'96131]
I'll keep looking, I guess.
[/quote]

I beleive that SyncBack (a Windows program) will check to see if there are changes to a file. This should help you determine if a file may be corrupt. I do not believe that it is a hard program to use.

--Ken
 
[quote author=Mark Sirota link=topic=975'.msg66'57#msg66'57 date=1273'9393']
Denis, if it'll work for you, I can share the relevant part of my script. The meat of it is:
Code:
diff -rq /Volumes/Photos-1/Pictures/Originals /Volumes/Photos-2/Pictures/Originals
It sends any the output to me by e-mail. I've been meaning to have it do something fancier with the output -- there isn't any unless there's a problem, so I don't want to risk losing it in my e-mail.
[/quote]
Thanks much Mark! Looks much like the equivalent Linux command! I now see much similarity between Mac OS and Linux... Do most or all Linux commands work? Or... can you point me where to look to learn all this for my Mac?

Yes this will work for me if it can compare from an HFS volume to a FAT volume. Waiting for the relevant script and how to do to avoid typing it in terminal each time I need it. :icon_neutral:
 
Linux is an operating system that is designed to look like Unix. MacOS X is based in part on BSD Unix, and is fully POSIX-compliant. Where Linux is also POSIX-compliant, the commands will be equivalent.

In looking over my script, I'm hesitant to share it because it's a bit ugly, is tightly tailored to my environment, and risks causing more confusion than it clears up. It does a whole lot of other stuff -- if there are discrepancies between the two main copies, it compares to a third copy, but the third copy is on my laptop and is incomplete, so most of the code involves handling that case.

If you're looking to run from cron, start with "man cron" and "man crontab". But it looks like the modern way to do this is with launchctl; if you're just learning this from scratch, you're probably better off figuring out how to do it with launchctl.
 
The file system type shouldn't matter when using diff as the system call that reads from the file system returns the file without any filesystem metadata.
 
Status
Not open for further replies.
Back
Top