4 ms·

If one can download a percentage for free each month - 5% in this case, and the price of storage is dirt-cheap, then couldn't one just dump empty blocks in unti

If one can download a percentage for free each month - 5% in this case, and the price of storage is dirt-cheap, then couldn't one just dump empty blocks in until the amount desired for retrieval falls under the 5% limit? In this case, if one wants to retrieve 63.3 GB, uploading 1202.7 GB more for a total of 1266 GB, 63.3 GB of which represents just under 5%. There's no cost for data transfer in and the monthly cost at $0.007/GB would be just $8.87. And that's just for the one month because everything wanted would be coming out the same month. Has anyone tried this or know of a gotcha that would exclude this? And I realize that for the OP's situation, it wouldn't have mattered since he thought he was going to get charged a fraction of this.

wallacoloo 11y ago
I believe that objects cannot be transitioned into the Glacier Storage class until 60 days after their upload date. So at the very least, you will have 60 days of latency (and thus ~$20 of monthly fees) before you can extract your other data. Additionally, I wouldn't be surprised if the 5% is also based on a storage measurement that is pro-rated for the month. So I would let the 1200 GB of data sit in Glacier Storage for another month before extracting anything, just to be (more) safe.
rakoo 11y ago
That's possibly a way to trick the system, but storing 63.3 GB in S3 would cost OP less than 2$ for standard availability and less than 1$ for reduced availability, not counting request costs (which are not as surprising as the hidden costs in question here). At this scale you should just store it in S3 and be done with it.
- captain_jamira 11y ago
  sure, I wasn't intending to suggest this as a good premeditated maneuver, but there are probably other individuals out there who have found themselves in a similar position to the OP and considering their predicament. If you've gotten into Glacier for the wrong reason, you may already be in the trap, and you can quickly rip yourself free and take a bunch of skin, spend almost 2 years ever so gently prying yourself free, or maybe a third way. That's my angle here. Also, traps don't have to be laid for someone to feel like he's in one, so I'm not putting that on AWS. The cheapest way out seems to be to just grab 5%/month over 20 months, but that's a lot of sustained effort and contact with the service. So I could see a trick like this as a potential middle ground, at three months and ~$30 according to previous comment's details.

wallacoloo 11y ago

I believe that objects cannot be transitioned into the Glacier Storage class until 60 days after their upload date. So at the very least, you will have 60 days of latency (and thus ~$20 of monthly fees) before you can extract your other data. Additionally, I wouldn't be surprised if the 5% is also based on a storage measurement that is pro-rated for the month. So I would let the 1200 GB of data sit in Glacier Storage for another month before extracting anything, just to be (more) safe.

rakoo 11y ago

That's possibly a way to trick the system, but storing 63.3 GB in S3 would cost OP less than 2$ for standard availability and less than 1$ for reduced availability, not counting request costs (which are not as surprising as the hidden costs in question here). At this scale you should just store it in S3 and be done with it.

captain_jamira 11y ago
sure, I wasn't intending to suggest this as a good premeditated maneuver, but there are probably other individuals out there who have found themselves in a similar position to the OP and considering their predicament. If you've gotten into Glacier for the wrong reason, you may already be in the trap, and you can quickly rip yourself free and take a bunch of skin, spend almost 2 years ever so gently prying yourself free, or maybe a third way. That's my angle here. Also, traps don't have to be laid for someone to feel like he's in one, so I'm not putting that on AWS. The cheapest way out seems to be to just grab 5%/month over 20 months, but that's a lot of sustained effort and contact with the service. So I could see a trick like this as a potential middle ground, at three months and ~$30 according to previous comment's details.