File Backup

Posted by: Mr Underhill on 13 May 2012

Since I bought the Drobo, having decided to rip my DVDs, 352 down so far, I have been deciding HOW to backup all my audio & video files.

The NAS gives me resilience - although I have read some disturbing reports on the Drobo nothing wrong so far, but I would HATE to EVER have to do this again.

In days gone by this was easy for me, two external HDDs hitched to an old Linux server with Rsync copying everything on Disk 1 to Disk 2 and I was sorted, but how do I work this now?

I cannot simply take a mirror copy of my NAS, not without buying a second NAS with 6TB of resilient storage, and growing!

I am sure I cannot be the only person in this position, and you may have come up with a better solution. Certainly I fervently wish the Drobo FS had one or more USB ports, it would make the whole thing a lot more straight forward, but it doesn't.

So ......



I decided to make the best use of the equipment I have to hand:

Two (2) 2TB HDDs,
One (1) 1TB HDD,
One (1) 500GB HDD,
One (1) old Dell laptop, running XP sp3.

The Dell laptop I set up a couple of years ago to rip my DVD-As, which I still occasionally buy when I see them at the right price. This now has Serviio DLNA Server installed - for streaming content from the internet & the Drobo.

The Drobo is built on Linux, and I have installed Rsync & NFS.

Traditionally I have hosted my files on two external HDDs, and kept another copy at a friend's as my off-site backup (OSB).

Step 1: Secure What I've Got.
I copied all my music files, WAV & flac, and photos onto the 1TB HDD and gave this to my friend = OSB.
I recovered my 500GB HDD, my old OSB HDD.

Step 2: Set up HDDs
Reformatted 500GB HDD & one 2TB HDD.
Both HDDs attached to Dell Laptop, kept in loft and running VNC.

Step 3: Incremental Backup ....er..


OK, in the good old days all my files would have EASILY fitted onto a 2TB disk, no longer. The video files are already 1.5TB, and growing daily - actually feels more like grinding daily.

My 2TB HDD will no longer cope on its own.

I used to be able to simply maintain a copy of what I had in two places, but now I need to split it.

Having secured my older files in my OSB I do not want to keep copying these.

What I want to do is simply copy my NEWER files to two different locations:

1. My video files to the 2TB disk; and
2. My audio files to the 500GB disk.

As the laptop is running XP my original thought was xcopy. This was scuppered as the command cannot work with shares onto the Linux based Drobo, can't cope with the file datetime stamps ...bugger.

My next though was to load up cygwin, this is a framework for adding unix commands onto a windows box. My hope was to then use Rsync to copy the newer files from the Drobo to the USB HDDs, there is a handy switch to allow this:

rsync --files-from=<(find /../src -atime +6 -print) src/ dest/

I read this, thought GREAT, and failed to get it working as described. Having played for a while I decided to dodge the issue by using a chain of commands, nice having unix and a windows box! In this way I could simply troubleshoot any issues on a command by command basis.

What I came up with was this:
find /cygdrive/w -type f -mtime -1 -print | while read file; do mkdir -p /cygdrive/f/BU/"`dirname "$file"`" && cp -an  "$file"  /cygdrive/f/BU"$file"; done

For those of you who are not familiar with BASH I'll break this down briefly. In essence this finds a list of files based on age (-mtime), and feeds this into a loop (using the pipe '|').

The loop first looks to see if the target directory exists (dirname), if it doesn't it creates it. It then copies the file into that directory - UNLESS in already exists, in which case it simply moves to the next file in the list (cp -an).

In cygwin all the local drives are mapped via the cygdrive directory:
/cygdrive/w = my video drive (W:\)
/cygdrive/f/BU/ = backup directory on my 2TB USB HDD (F:\).

find
-mtime = file modification time & -1 means within the last day.

This list of files is printed and fed one at a time into a loop as the variable 'file'.

mkdir
If the directory doesn't exist it is made, the directory path is extracted from $file - the earlier created 'file' variable.

The directory path is appended onto /cygdrive/f/BU/.

The path has to be surrounded by speech marks to cope with spaces. As Naim picked windows as its OS for its servers spaces abound, hence: "`dirname "$file"`".

This then leads into the copy command, which will now succeed as the target directory exists.



I have set up three such commands: 1. For my NS01 share; 2. One for my NS01 store, and; 3. One for my video files, all hosted on my NAS.

These are Scheduled Tasks within Win XP running in the early hours.

Once the HDDs are full they will go to my offsite BU location, and new ones purchased.


Hope this may help, or give you ideas.

M






Posted on: 13 May 2012 by Tog
Sounds complicated - but the golden rule is that it has to work for you - to the extent that you backup regularly. My music is backed up to two 1 TB drives twice a month ( drives kept in different places) - easy with Vortexbox. Video via Plex is stored separately and I'm looking for an equally simple backup routine. Tog
Posted on: 13 May 2012 by rjstaines
Originally Posted by Mr Underhill:

Hope this may help, or give you ideas.

M


Hmmmm... just wondering how honest to be here, Mr H....

 

I remember pondering the likes of what you're doing, then thinking "Nah sod it, I'll buy another NAS drive and stick it in the shed"  ..and that's what I did.

Posted on: 13 May 2012 by Mr Underhill

Hi RJ,

 

Yep - that is a straightforward answer .......unless you get burgled and they nick both boxes.

 

The thing I liked about this was making the most of some bits I had lying around.

 

The most painful part of this was setting up the Scheduled Task under Win XP!

 

Whatever happens I want to make sure I never have to re-rip my CDs, LPs or DVDs.

 

 

Hi Tog,

 

Yes - and two 2TB disks used to do for me - no longer.

 

As you say - '...golden rule ... you backup regularly'. I have ended up helping a number of colleagues who had the best of intentions.

 

M

 

Posted on: 13 May 2012 by McGhie

I signed up to an offline backup (cloud) service last month.  Unlimited size backup from a single machine for <$140 for 4 years (<$3 per month).  I back everything up to my server (in real time) and then back the server up to the cloud (as well as using mirroring on server and running twice daily local backup of server).

 

The only problem with the cloud backup is that it is dependent on broadband upload speed, which isn't great for me.  Still, I'm managing 10GB per day and am 80% of the way there in backing up a modest 500GB of data (looks like it'll end up taking less than two months.

 

I'm using CrashPlan, which provides quite a few options for backing up, and if you wanted to use it to back up a large amount of data (at my upload speed it'll take about 3 months per TB) you could prioritise it so it does your most important data first.

 

Just a thought...

 

Cheers

Ian

Posted on: 13 May 2012 by Bart
Originally Posted by McGhie:

I signed up to an offline backup (cloud) service last month.  Unlimited size backup from a single machine for <$140 for 4 years (<$3 per month).  I back everything up to my server (in real time) and then back the server up to the cloud (as well as using mirroring on server and running twice daily local backup of server).

 

The only problem with the cloud backup is that it is dependent on broadband upload speed, which isn't great for me.  Still, I'm managing 10GB per day and am 80% of the way there in backing up a modest 500GB of data (looks like it'll end up taking less than two months.

 

I'm using CrashPlan, which provides quite a few options for backing up, and if you wanted to use it to back up a large amount of data (at my upload speed it'll take about 3 months per TB) you could prioritise it so it does your most important data first.

 

Just a thought...

 

Cheers

Ian

And if you needed to recover, do you have an option faster than the same 10gb/day downloaded?

Posted on: 13 May 2012 by Zinger
I have the same dilemma as you do. I have a few harddrives lying around, but that does is not the best way to back things up as things will be in pieces. But neither am I ready to drop another sum of money for another NAS + 4 x 2T drives. I just run RAID5 on my machine, and pray that the NAS can rebuild fast enough if one drive fails... If something physically happens to the drive, then I guess that's it ...
Posted on: 14 May 2012 by trickydickie

Crashplan also allows you to backup to a local drive or you can backup to a drive owned by you at a different location, i.e. a friends house.

 

Richard

Posted on: 14 May 2012 by McGhie
Hi Bart Four things to consider when it comes to download (as opposed to upload): 1. I'm hoping to never need to download anything as I have a copy on my PC, a copy on the server (RAID1), and a backup of the server, so the data will survive most catastrophic scenarios (I'm buying peace of mind and mitigating the risk to my data of fire/theft, though the backup is also incremental, so CrashPlan allows me to restore previous versions of particular files, in case I realise too late that I've deleted or overwritten something). 2. Broadband download speeds are significantly greater than upload speeds (many times faster) - it'll take a fraction of the time to download. 3. If, unlike me, you live in North America, CrashPlan provides a service where you can request that, for a fee, they ship your data to you on a disk, which you can then buy or return. (They also provide a 'seeding' service in NA - you can send your data on a drive to avoid the initial huge upload. Maybe these services will be offered outside of NA some time). 4. CrashPlan, as well as providing paid cloud backup plans, provides free functionality to backup to another machine (rather than their cloud) over the Internet (e.g. to a drive on a friend's PC), which gets around the download time, as you could get hold of the drive by some other means. I'm happy to work on the basis that in the worst case scenario it might take 10 days to recover my 'lost' data. Cheers Ian
Posted on: 14 May 2012 by Tog
For music (flac) 2 x 1 TB drives is ample for the time being. The Video mkv files generated for Plex will require at least 2 x3TB as backup and a RAID array for playback. Tog
Posted on: 14 May 2012 by Phil Harris

I run a 27Tb Windows Home Server based NAS (running "DriveBender" to do flexible drive pooling) as my main "media" storage system at home ...

 

...and have a second 27Tb Windows Home Server based NAS (running StableBit "DrivePool" to do flexible drive pooling - I use two different solutions in case one has a fatal flaw) which is used solely to back up the first server.

 

Re-ripping all my CDs, DVDs and BluRays is something that I've done once too often and don't wish to do again ... even doing a full backup of my main server takes a month! (Thankfully the weekly incrementals generally take less than 20 minutes.)

 

For personal documents and photos I have a reciprocal arrangement with a friend whereby we have a small separate NAS each and I have 2Tb on his NAS and he has 2Tb on my NAS that we keep synchronised between us in case of "disasters". 

 

Phil

Posted on: 14 May 2012 by Mr Underhill

McGhie:

offline backup (cloud) service


Hi Ian,

I am deeply suspicious of using 'The Cloud'. Ultimately my files must be sitting on metal somewhere, I would be reading the policies carefully - and encrypting everything that is at all personal.

In this case what I am handling is rips of CDs / LPS and DVDs, other peoples copyrighted content. As you know we have no reasonable use provisions in the UK. I believe that as I am ripping my media for use within my own home that FACT will not be interested in me - I spend a lot of money buying my content. However, if I was to upload this to a site on the Internet ......? I appreciate this is NOT aimed at sharing, but in their position I would take a very dim view of this.

One thing about a backup regime is that you should periodically check that the files you are backing up CAN be used to recover. As I am sure you know Microsoft and Amazon have both had major outages on their services - in your position I would drag a few files back on a monthly basis, just to have peace of mind.


Zinger:

This might give you an automated way to get some more use out of those HDDs you have lying around?

 

trickydickie:

'..or you can backup to a drive owned by you at a different location.'


Which is what I do. The code backs up the incremental differences over the previous 24 hours to external HDDs, as these fill they get taken to a friend.

My issue is that I wanted to automate this, but I couldn't base the differences on the whole file structure, as that has grown too big.

 

AllenB:

'...the biggest chunk of your data are your ripped DVD's'


Hi Allen,

At the moment my CDs = 300GB
HiDef Files (inc LP rips) = 500GB
DVD = 1.5TB

So yes, although that is simply because ripping LPs is such a long process, and I've run out of steam for the moment. DVD is easy to productionise.

I do keep ALL my original discs, this is how I would prove my 'ownership'.

The time this is taking is significant, I will do my damnedest not to do it again!

 

Tog:

I have ripped about 20% of my DVDs to date - so I am heading towards 8TB just for these, I have barely scratched the surface of my LPs, 200 or so.

 

Phil:

'I run a 27Tb Windows Home Server based NAS ....and have a second 27Tb ....I have a reciprocal arrangement with a friend'

'Re-ripping all my CDs, DVDs and BluRays is something [I] don't wish to do again'

 

Which is broadly what I am doing, but want to avoid buying a second NAS. Hey, let's face it, I'm just cheap.



Using cygwin in combination with Scheduled Tasking was an interesting journey.


M

Posted on: 14 May 2012 by McGhie
Interesting points. CrashPlan uses strong (448-bit Blowfish) encryption, so I'm relatively comfortable that my personal data is safe and that my rips (of CDs that I have paid good money for) shouldn't attract unwanted attention. I certainly wouldn't want my data shared with anyone. I will check that I can recover a sample of files once the upload is complete. It is a worst case scenario fallback plan but it would be pretty pointless paying for and running the thing if it was not possible to get files back from it if required. Cheers Ian