I came into some Amazon gift certificates I’m figuring out how to spend (aka burning a hole in my pocket), and am considering a second Drobo FS for automated secondary backup. The idea is that I keep all my critical files on the Drobo, and I could rsync or use some other Drobo app to auto sync to the other drobo. Then I just keep a copy on me for offsite backup. Thoughts? What sync application would be a good choice?
Given the Drobo premium it seems a little over the top, but why not?
Although I am probably the biggest “apologist” for DRI around here, I have to ask: have you considered the Synology DS1511+?
I very much love my FS, but if we are to approach this from a data security mindset, you are probably better off having hardware from different manufacturers.
Personally, the absolutely essential files on my FS are backed up on an old Linkstation Live v2, which sits on my desk at the university. (Funny fact: the Linkstation Live v2 and the FS have the same family of CPU, the FS is just clocked at 800 MHz while the LSLive is only 200 MHz)
The synchronization method is rsync over ssh. Nothing beats that for openness. It may be a bit tricky to setup, but once it is done it will basically run forever. And if you feel like overdoing it, the sky is the limit (e.g., email/sms notifications, bandwidth throttling, dropbox-like behavior and so on…).
Ricardo has a good point. for simplicity and sanity to keep track of things you might have an urge to keep them truly identical. You could end up doing a firmware update to both at once, creating a problem on your working and Hail Mary backup copy. Or you might just change some configuration that causes a problem. It’s a single point of failure that you might want to avoid.
I backup my Drobo V2 to bare drives that I put in BlacX docks. Although my Drobo and the docks are connected to a Win 7 machine acting as a NAS server, you can do the same via gigabit ethernet with the docks attached to your working machine. I avoided the cost of another Drobo and all those potential single points of failure. Plus, for backup purposes, I get a warm and fuzzy backing up to bare drives that my native OS can read. Push come to shove I can mount those drives in any machine.
Or for about $300 or so you could build your own Linux or FreeNAS Nas and have some fun with that. Then you could have horse races at night between the Drobo FS and the DIY NAS for added entertainment, which is truly priceless
Getting two different platforms is an interesting idea, but what is it really protecting you against, systemic problems in the platform? Consider the big cloud centers, do they spread data across hardware vendors? I’m pretty sure they don’t. That makes me wonder about how safe cloud data is. Amazon S3 is pretty safe from what I see, but this data center Apple is building (for who knows what reason), ‘all’ it takes is a fire to bring that down.
Anyhow that Synology is pretty sweet, and pretty expensive. Neil, the reason I got a Drobo is because I tried building my own Linux NAS, ended up spending as much money and it didn’t work very well. I presently have disks on my other computers that I run DroboCopy from, but DroboCopy is particularly stupid. If the disk doesn’t happen to be mounted (such as when waking it from sleep) it will bitch and not do it. I want something fully automated and easy to set up.
The other idea I had was to get a Mac Mini (or repurpose the one my kid is using) to be the rsync server from the Drobo.
This is very hard to tell, since this is kind of secret data. We do know for a fact that Google uses HDDs from all manufacturers, since they released a study a few years ago saying that in their experience paying extra dor fancy “enterprise-grade” drives is bollocks. Keep in mind that they do not treat their servers so well, since Google is known to run everything a few degrees too hot to save on cooling. But you could say that it is most likely that all the servers within a single datacenter are identical.
Another thing is what NeilR said: in such datacenters you’ll never ever see them deploy the latest firmware on all their servers at the same time, since if it has a new bug it won’t affect everything. Usually they take a few nodes out, make the update, run them under supervision for a few days/weeks/months to make sure everything is still ok, and then deploy to the rest of the machines.
If you have only two devices at home, however, the temptation is too big to update both of them at the same time and suddenly lose everything. By having devices from different vendors you get free of this temptation since vendors rarely release updates on the same schedule.
About cloud providers: data can be made as available and reliably stored as you wish. The problem is that availability and reliability cost money. Exponentially more, as the requirements increase. The whole Amazon debacle a few weeks ago was because people used the cheapest cloud storage from Amazon, and wanted the reliability of their most expensive offering… and then sh*t happened.
That seems like a pretty sweet idea. You could even have the Drobo wake the Mac Mini when a backup is about to start.
Clouds are just another single point of failure. If all your data is in one cloud, you need another as backup, for all the same reasons you need a backup for your Drobo. The only difference is that there are IT professionals running the cloud, which should decrease the chances of failure. But the probability of failure can only asymptotically approach zero. It never hits the zero line. You just need to decide how close to zero you need to be.
I didn’t think about the Amazon cloud problem as a “debacle” because no data was lost. It was just a short term inconvenience of sorts. The rest of the world apparently did, or talked it up as such. Cloud providers have lost their customer’s data, in particular consumer cloud backup providers such as Mozy (not Mozy, but someone like it comes to mind but can’t remember who). I would never leave my only copy of my data with Mozy or Carbonite. No one can get that probability to touch zero and stay there forever.
I keep 3 additionalcopies of all the data on my Drobo V2. And a special 4th copy of my photo images to protect against corruption even though I shoot raw and therefore don’t need a backup copy to revert to the original, such as I would with JPGs.
I do that partly because even I’m not stupid enough to rip through the 4th copy of my data before pausing to catch my breath and think about why I ripped through the first 3. That’s the theory, anyway. And for other reasons, other various and sometimes obscure single points of failure.
[quote=“toolusr, post:4, topic:2616”]
Getting two different platforms is an interesting idea, but what is it really protecting you against, systemic problems in the platform? Consider the big cloud centers, do they spread data across hardware vendors? I’m pretty sure they don’t. [/quote]
Speaking as a datacenter/network architect, I can say the good ones do. That doesn’t mean every machine is different, but you at least have clusters of different machines for exactly that reason. Multiple firewall layers from different vendors. Same thing for OS if possible.
Back on the topic at hand, I completely agree with Ricardo’s suggestion. Since syncing data across the systems won’t care what the underlying hardware is, you might as well get some diversity. And quite frankly, you might end up liking the Synology more. Sometimes I feel that it just can’t be worse than the DroboFS.
Hmmm, despite being a software architect I don’t want to mess around with scripts and whatnot, too busy doing other stuff. I’ll admit that despite the hefty price tag this Synology looks pretty compelling. Tons of features and the ability to sleep and wake up - I’d love to sleep the NAS at night.
Heh heh … I ordered the 1515+, you a bad, bad influence on me
The more I read the features the more I realized it was exactly what I was looking for, then got a demo from a friend who has one, the software is pretty amazing. Not sure I’ll even need the Drobo now.
One of the reasons I bought the Drobo was for promised cloud backup, outside your firewall access etc. OxygenCloud beta didn’t impress me, and the feature is unavailable to anybody but a paid business account now (supposedly will generally available later this year but who knows). Actually this is what prompted me to post this thread. But it seems that Synology supports Amazon S3 which is good enough for me.
I’ll see but I think I should have bought the Synology in the first place. USB backup, eSata expansion, scheduled wake up/sleep, extensive software platform, grrrrrr