Unlimited storage on Amazon cloud drive for $5

Hey guys, I read an article today which says Amazon is making their cloud drive unlimited storage for 1 year available for $5. Seems like a crappy interface, but for $5 for unlimited storage, who cares.

http://mashable.com/2015/11/27/amazon-unlimited-cloud-storage-deal/#DmOVQBpWSaqs

Regards,
John

Same price as the Raspberry PI zero. 5 Bux :wink:

So, I’m not exactly an rPI fan, by far. But for 5 bux . . . Or even free if you live in the UK, and have the ability to buy MagPI #40. As one comes in a packet attached to this months dead tree copy of MagPI.

First one's free: you upload terabytes of data, then your year is up, and what're you gonna do?

By then there will be even cheaper options available :wink:

Regards,
John

I saw this yesterday. Amazing.

Regards,
John

I saw this yesterday. Amazing.

Regards,
John

First one’s free: you upload terabytes of data, then your year is up, and what’re you gonna do?

Look over at the terabytes of data I have stored in my room. . . I never understood cloud storage.

My guess is you do normal backups of all your important work; however, what if you have a fire, theft, or some other disaster, which will destroy all your backups as well. Hence the need for offsite storage. Now unless you are storing your backup tapes/disks offsite, cloud storage starts to make sense.

Regards,
John

My offsite storage is in my garage which, fortunately, happens to be
50 metres or more from the house.

Our broadband isn't broad enough to make cloud backup remotely sensible.

I like my backup-backup nas… The base board is bricked… So the data is securely saved… :wink:

My guess is you do normal backups of all your important work; however, what if you have a fire, theft, or some other disaster, which will destroy all your backups as well. Hence the need for offsite storage. Now unless you are storing your backup tapes/disks offsite, cloud storage starts to make sense.

This is the “excuse” if everyone using cloud storage. Simple fact is, there is no data I have stored that is that important. All of it can be replaced. Pictures, code, whatever.

Not to mention a fire is very unlikely, but if there were one, if I were not able to put it out, it would likely kill me anyhow. Rendering my data moot. Theft ? well lets just say a thief would very likely have few dogs on him, as well as a couple bullet holes. Someone is always here.

A few years ago I was backing up all my data to a raid6 server. My thinking was any two disks can fail simultaneously and I still would not loose data. Unknown to me, I was using an Intel RAID controller that had a firmware bug and it trashed all my disks and I lost about 6 months of work. Now I do my backups with belts and braces so nothing like that can ever happen again. I now have multiple RAID servers which mirror each other and no one machine has components in common with the other machines. To me, cloud backup was just another redundant offsite backup, but the Amazon tools are horrible and the service hangs for no reason. Needless to say after fighting this all last night, I decided to abandon the Amazon cloud drive.

Regards,
John

For my usage, RAID is useless. Better to use separate disks, and rsync. As most data does not need to be redundant, and you get more storage that way, with very little to go wrong.

Yeah, but rsync only gives you a snapshot and not a history of your backup. When I really mess up, I want to go back to the state of my machine 15 minutes ago, or two days ago. This has saved me a lot of head scratching, trying to find out where I messed up. I really like the way timemachine works on the MAC. I can add as many disks as I want to the timemachine and it just sequences the backups between each of these disks. This happens every 15 minutes. If I loose a disk, then I loose 15 minutes of work.

I also have a SMART monitor that keeps an eye on the condition of these disks, looking potential disk failures.

For Linux, I haven’t found anything equivalent to timemachine. BackinTime and the likes all try to do the same thing, but after a few weeks they all end up using tons of CPU time and become so slow. In essence, these solutions use rsync to create snapshots each hour and then use hard links to eliminate duplicating unchanged files. Currently I’m using Crashplan which performs well, but I don’t get the redundancy.

Regards,
John

Yeah, but rsync only gives you a snapshot and not a history of your backup. When I really mess up, I want to go back to the state of my machine 15 minutes ago, or two days ago. This has saved me a lot of head scratching, trying to find out where I messed up. I really like the way timemachine works on the MAC. I can add as many disks as I want to the timemachine and it just sequences the backups between each of these disks. This happens every 15 minutes. If I loose a disk, then I loose 15 minutes of work.

  • a) rsync for files that do not need incremental backup. e.g. pictures etc.
  • b) git for files that do.
  • c) dd, and /or tar for the whole OS file system.
  • d) Fancy scripts or executable to put it all together. If wanted.

a) rsync for files that do not need incremental backup. e.g. pictures etc.

And by this of course I mean: Perhaps you have a whole directory for pictures, and those that have already been backed up using rsync wont need it done again. But if you add new files in a new directory . . .then let git handle paths that do need incremental backups. Daily, hourly, whatever. Usually source, but who says that a local repo has to be source ?

I've never seen the point of RAID in a domestic situation. It's more
for providing secure, always available, disk storage, not for
providing backup. Apart from anything else if the computer catches
fire or is destroyed by some other means (or stolen) then all your
data is gone, no matter how redundant the strorage was.

Some sort of off-site, or at the very least 'off system' backup is
what's needed to preserve data.

I use an rsync based incremental backup system (I wrote it myself
having used rsnapshot for a while, rsnapshot is OK but I think it's
too complex).

I do hourly incremental backups locally to another disk on my main
machine and I do daily incremental backups to a remote machine. The
daily remote backups get thinned out as they get older so there are
daily backups for the last month, then monthly ones for 12 months,
then yearly ones.

That makes perfect sense. BTW, the only purpose of a RAID backup is to prevent a single point of failure (like a disk failure) resulting in lost backups.

One thing to pay attention to is the MTBF numbers for disks. I was a firm believer in Seagate Barracuda disk until I had a whole number of them fail over a few months. Speaking Seagate tech support, they explained that the SMART data on these disks showed they had more than the 3,000 hours MTBF and hence I should have expected them to fail. I couldn’t believe what they told me; running their disks 24 hours/day, they expected failures in 1/3 of a year. They were right, look at the SMART data on Seagate disks and you will see read write errors in the 10’s of thousands or more.

After that I use Western Digital RED disks which are designed for 24/7 NAS applications. Looking at the disk SMART data, I see 0 read/write errors.

Regards,
John

. . .the only purpose of a RAID backup is to prevent a single point of failure (like a disk failure) resulting in lost backups.

You do not need a RAID array to prevent a single point of failure. You take those 3+ disks, put them in 3 different machines. Or even in the same machine as single drives. Same difference, only less wear and tear on the drives, more cost effective, and perhaps a small amount slower as singles.

In the field you’ll likely not run into any RAID 5/6 arrays. At least for corporate storage. You’re more likely to see RAID10, or RAID0 + 1. Because there is nothing faster than striping disks, and RAID1 does not have an impact on performance if set up correctly. RAID5/6 is just a way for the home user to feel all warm and fuzzy . . and literally feed the companies who offer the hardware for such arrays. Be it controllers, or “special” hard drives . . . special software, chipsets with BS built in RAID( software ).

I still use Seagate drives(nothing but), and have no issues. Why ? Probably because I do not run RAID. RAID is notorious for being hard on drives. Especially RAID 5/6. I will admit, that Seagate’s reputation has gone into the toilette in the last 8 or so years. All their drives used to be lifetime warranty. Now days I think they give 3 years . . . not even as good as WD, or even Samsung SSDs . . .

Anyway, seriously. Unless you’re running a server that sees thousands+ of transactions a day. You don’t need RAID. But hey, don’t pay attention to me. . .