Wijsmullerbros.nl

By all means let's be open-minded, but not so open-minded that our brains drop out!

— Richard Dawkins

Aug
30

We finally decided to buy a backup device, a so called "NAS" (network attached storage). The reason is that we have been building up a collection of photo's, clips, documents and other stuff that we just don't want to lose anymore.

In the past, I used to store files on my laptop, my wife's laptop, on an external drive, and now and then burn a dvd with the most valuable data. The problem is that now it's really becoming too much data and I'm just not that fond anymore of spending my time doing this tiresome copying all the time.

The Synology solution has room (in our case) for 1500 gigabytes, far more then we can collect in the coming years :D. Better yet, we physically have twice the amount, but half is lost because of internal duplication. This is meant to be able to recover in case of storage failure such as a harddisk "crashing" which happens a lot! (let this be a warning). This system is called RAID1 by the way.

This specific NAS is awesome, it downloads, shows photos, plays music, makes coffee and much more! (ok, maybe it doesn't really make coffee).
Up until now, the only annoyance was some noise coming from the fan, but that turned out to be a big fly stuck in the fan which was easily removed using a screwdriver.

I'll probably be babbling some more about this great machine in the near future, so run and hide soon before a start again ;-)

3 comments

Wicher

But do you keep an offsite backup? I use RAID1 only for availibility purposes, for services that are time-critical, such as a mail server.
For media, instead of RAID1, a saner approach is to use the 2n available storage space to store a copy offsite. If done intelligently, it does not only protect you from natural disasters but also from stupidity such as making typos when rm -rf-ing.
On the subject of natural disasters: If I recall correctly, you live on a floodplain ;-)

braam

The problems for me currently with offsite backups are stability, costs, security and durability. Or to put it differently: Will my data be available when I most need it, safe from preying eyes, kept in storage forever and affordable?

I'm afraid that all the stuff I've seen so far doesn't quite meet these requirements.

And I don't live in a flood plain (though I do live below the river water level).

Wicher

I'm still pondering solutions, too.

I had a look at LAFS-Tahoe which is great but not bandwidth-efficient (my presumption is that it's file-based).
I want to transfer blocks, and only the ones that do not exist on the other side yet.

Currently I'm investigating lessfs with the TokyoDB TokyoCabinet backend: it does deduplication, compression, encryption and with TokyoCabinet running in replication mode, it does automatic distribution of new and changed blocks.

Having an offsite backup becomes as simple as running a TokyoCabinet slave (through Tokyo Tyrant?) on one of your friends servers. You need about 3n worst case storage space (original data included), compression and deduplication should be able to trim that down to about 2.5n. But the upside is that you'd transfer the absolute minimum amount of data, much less than rsync. And it's not the storage space that's at a premium, it's your and your friend's bandwidth ;-)