Thawing/freezing timescales for objects in AWS Glacier Flexible Storage

Question

I have a number of objects that I wish to restore from Glacier Flexible Storage (GFS) and am trying to understand how the thawing/refreezing process works. As I understand it, I can provide a parameter, say 3 days, which means that the object will stay thawed for three days before automatically freezing again.

In my case the usage time could be fairly quick or it could last a day or two, so normally 3 days would be a good choice. However, there may be a time when I want to restore and then quickly use the object, but then a day later I want to do the same again on this object. In this case, the second restore attempt would fail as the object is already restored.

The simple answer to this might be to check the file status (thawed/frozen) before using, so if it was thawed, I would just skip the restore and use. The problem here comes if I want to use the data real quick on day one but then come back late on day two and will need the object for some time. In this case, I might start using the object only for it to refreeze under me.

Is there a way to trigger the re-freezing of the object once you know you don't need it any more, thereby syncing access, and does this have a financial cost?

Answer

Hey component_10,
I understand that you would like to find a way to automate the retrieval of the archived object. I presume you are implying that the issue is that it is hard to gauge how long you will need the thaw period for a certain batch of objects in Glacier Flexible Storage? It seems there might be a bit of a misunderstanding regarding how glacier works. Glaciers main purpose is for data archiving. If you wish to keep GFS and are still looking for faster retrieval requests, please look more into S3 expedited requests (more costly). For reference, in your case, "GFS is used for archives where portions of the data might need to be retrieved in minutes. Data stored in the S3 Glacier Flexible Retrieval storage class can be accessed in as little as 1-5 minutes by using Expedited retrieval. You can also request free Bulk retrievals in up to 5-12 hours" [4]. If this is not what you are looking for as you wish for faster retrieval times, but want to keep costs at a minimal, you can take a look into S3 Glacier Instant Retrieval. It is not recommended to have items in Glacier if you are accessing them frequently within the span of a couple of days. Take a look at your access patterns and align them with the right S3 class to fit your needs whilst balancing costs [5]. For example, you can consider swapping to another S3 storage plan such as One-Zone (for costs) S3 Infrequent Access (IA) in order to retrieve on demand, while not necessarily knowing the time frame. The costs for these plans follow this document [1]. For a more cost-optimized path, you can consider creating a S3 bucket lifecycle policy to transition objects that haven’t been accessed in a while to another plan such as Glacier Deep Archive [2]. If this option is not necessarily applicable, please feel free to clarify the problem/response, and I can then re-adjust my answer. 
Thanks!
 
References:
[1] https://aws.amazon.com/s3/pricing/
[2] https://docs.aws.amazon.com/AmazonS3/latest/userguide/analytics-storage-class.html
[3] https://docs.aws.amazon.com/AmazonS3/latest/userguide/object-lifecycle-mgmt.html
[4] https://docs.aws.amazon.com/amazonglacier/latest/dev/introduction.html
[5] https://aws.amazon.com/s3/glacier/pricing/

Thawing/freezing timescales for objects in AWS Glacier Flexible Storage

Relevant content