[Digikam-users] Backup options?? probably off topic

[Photonoxx]

Le Thu, 18 Jul 2013 02:07:55 +0200, Simon Cropper
<simoncropper at fossworkflowguides.com> a écrit:

> On 17/07/13 12:34, Fabio wrote:
>> Hi All,
>>
>> I am not sure this is the proper channel but I could not find any other
>> digiKam forum
>>
>> Digikam strong area is organization and exporting to other places. Now,
>> I am looking for backup options. The easiest way is just to copy all
>> files from my local album to 2 external drives. Problem is Once I made
>> the first copy, how can I keep updating only photos I have been
>> changing. I am going back and editing old photos but no tin a very
>> organized way.
>>
>> Does digiKam provide backup options like incremental backup (copy only
>> what is new or changed)?
>>
>>
>> Thanks!
>> _______________________________________________
>> Digikam-users mailing list
>> Digikam-users at kde.org
>> https://mail.kde.org/mailman/listinfo/digikam-users
>
> Fabio,
>
> Very large image collections, as created by digiKam, creates a unique  
> scenario and so I think it is not that "off topic" to ask the question  
> on this forum.
>
> The way that digiKam works is to catalog all the images and associated  
> metadata in a database; usually stored in the top most directory of your  
> albums. Although uncommon this lists has reports of people trying to  
> recover their tags and metadata from corrupted databases as they have  
> not ensured that this information is stored elsewhere.
>
> DigiKam provides the ability to save metadata to associated XMP files  
> and to store metadata tags, copyright information and captions inside  
> the files (JPGs only). I would recommend you do this -- it generally  
> does not have a significant overhead. This option allows you, in the  
> circumstance where you database becomes corrupt to recreate your  
> database based on the information stored in these files.

Storing metadata works with some RAW files too. I use Digikam for this
almost without issues. I had two or three times some fright due to some
Exvi2 issue few years ago, but most files had been repaired.

XMP files could be a nice option too because it make much less MB to
synchronize/backup, I should test again since the feature seems mature
now, and for the story a "long" time ago backup for picture
only if the included metadatas had changed was not possible since the file
attribut "modify date" was not changed by Digikam, but it's fixed from a
long time now.

> The next issue is conventional backup. Albums are just directories so  
> are easily backed up using any of the packages mentioned (rsync, Unison,  
> BackupPC, grsync, Timemachine). Ideally you should have all your  
> electronic files backup, so asking the question is a little bit of a  
> concern :).
>
> To help...
> 1. Solutions are either GUI-based or Script-based. GUI-based solutions  
> are good to get you going. Script-based solutions require some level of  
> "programming" so you will need to take the time to read the manual,  
> construct a script and test it.
> 2. Ensure your backup creates a mirror of your original files. That way  
> if your primary disk dies you can just start using one of your backups.  
> Some incremental software bundles zip collections or compress the files  
> and you need to uncompress and reconstruct your original collection if  
> you wish to recover your files. This can be a nightmare!
> 3. You need to decide if you want some level of versioning. If you  
> mirror your collection and corrupt a file unknowingly you can have the  
> corrupt file copied to you backup resulting in the loss of the original  
> good file. In general, due to size of most people's collections people  
> don't create versions. If you use scripts however it is pretty easy to  
> make copies of your collection for each month. Hard disk space is cheap  
> and a script that copies originals to a safe place for a couple of  
> months are not that hard to make.

Versioning is a good question and YES, it take much much place on a hard
drive.

Since in my mind the main importance thing is my pictures integrity and
only after comes the metadata/treatments preservation, here what I do :

-I import my pictures with Rapid-photo-downloader which offers possibility
to rename and organize the pictures in folder depending of their metadatas
like digikam, but, unlike Digikam (except if I missed this new feature ?),
permit to set a backup import location. This way I potentially always have
my original picture (the better could be this backup location was a mirror
RAID)

- I try to backup my active working directory regularly with a sync
program.

>
> All that said, I will tell you what I use...
>
> On Windows I use RoboCopy. It is a command line program that can have  
> instructions kept in a batch file. The program is free from Microsoft.  
> The syntax is easy... for example...
>
> robocopy c:\images \\backup\images /mir /ns /R:2 /NP /NDL /Purge /tee
>
> I have tried a range of packages on Windows but the only software I  
> liked was AllwaySync -- a good application for small collections.
>
> http://allwaysync.com/
>
> *I now primarily use Linux*. All my backup is automated using rsync.  
> This package is very much the same as RoboCopy. You insert a series of  
> commands in a script file and run the script. My script creates a mirror  
> of my entire network, including my image collection. Again the syntax is  
> easy... another example...
>
> rsync -rltvz --stats --delete --ignore-errors /home/simon /media/backup
>
> In some situations, on Linux, where I want short term backup solutions I  
> use FreeFileSync. It is easy to use and to create simple backup options.
>
> http://freefilesync.sourceforge.net/

I was using FreeFileSync sometimes ago, but I was not sure it was
still maintain. It is a good news it was since it is really a simple
and efficient sync GUI programm with different sync mode.

>
> Remember...
> A. Backing up large collections over a network can take forever. Make  
> your initial backup by connecting your backup drive to your computer via  
> USB or actually plugging it in (internally). Then move it to a remote  
> location once the backups are completed. Incremental backups should not  
> be too much of a load on a normal network.
> B. All backups require you to be able to "see" the backups. If you  
> temporarily connect a USB drive this is not a problem for either Linux  
> or Windows. When you work with remote computers however it is usually  
> easier to mount the drives. On Windows this is done by mounting a  
> directory on your backup drive to a letter. This allows you to insert  
> your name and password once and have that mounted drive open whenever it  
> is available. On Linux, this is a little trickier. Mounting is done  
> using the mount command and requires a little effort and testing. The  
> exact syntax depends entirely on the distro (e.g. Ubuntu, RedHat), hard  
> disk format (e.g. ext3, ext4, NTFS, FAT32) and connection strategy (e.g.  
> NFS, CIFS).
>


-- 
Nicolas Boulesteix
Photographe chasseur de lueurs
http://www.photonoxx.fr