S3 Storage Class "Infrequent Access" Support

[mattbarb]

Hello, Any plan to include support for the Infrequent Access storage class of S3? I know you can choose Reduced Redundancy currently.

[superflexible]

Thanks, it will be considered!

[kquiring]

+1 on the request... storage costs on Amazon S3's "Infrequent Data" variant are less than half those of the standard S3 service. Supporting S3-IA would immediately result in considerable $avings for my application, so I'm hoping this can be added.

I'm new to Syncovery but really like it so far... it's very flexible in the control it offers, and it seems to be organized well. Thanks for the great app.

[Luc23]

I could be wrong, but I don't think this is possible. I believe that objects in S3 must reside in S3 Standard for 30 days before they can go to S3-IA. This can be set in the Lifecycle Policy of the bucket.

[kquiring]

I'm not sure what resource led me to conclude it was necessary. But re-looking into it now, this FAQ suggests either is possible. (See question 'How do I get my data into Standard - IA?')

https://aws.amazon.com/s3/faqs/#sia_anchor

But the lifecycle rule works nicely for me, so I retract my request. The only reason I'd imagine someone needing the PUT configuration is if it's unpredictable for a given piece of data within a bucket whether it needs to be IA or not. That isn't my situation... I simply configured the whole bucket. The rule didn't limit me to a 30-day transition window; you have some flexibility.

Thanks for the tip, Luc23.

[Luc23]

Yep, it does look like you can directly Put objects in S3-IA according to that FAQ. Looks like we both learned something, thanks!

One other thing to note that you may already know. I personally stopped using RRS entirely. It actually seems somewhat deprecated in AWS since S3-IA was announced. The thing is that going from RRS to any other form of storage (S3, S3-IA, or Glacier) requires a move of the object because its physically different storage. Where as going between the other three is just changing a flag on the backend. But I suppose I could be wrong on that too, so just FYI.

Update:

I just tried testing creating an S3-IA bucket through Cloudberry. And although the option is there to do it, it throws an error and doesn't seem to work. I think that might be because I just installed the newest version and it may want a reboot which I can't do right now. I then went in to the S3 Management console and found there was no option to create an S3-IA bucket directly. Using hte Lifecycle rules, you can tell it any number of days to flag the objects as S3-IA, but there is this disclaimer right below it:
Standard - Infrequent Access has a 30-day minimum retention period and a 128KB minimum object size. Lifecycle policy will not transition objects that are less than 128KB. Refer here to learn more about Standard - Infrequent Access.

This may just be a recent transition. I've seen many times with AWS where they change something (make something better), but take a bit to update their documentation to reflect it. I'm curious to see what your findings are on whether a new object shows in S3-IA in less than 30 days.

Update to my edit:

Lol... I just actually tried moving the days to less than 30 and you get a big red message on the screen not allowing it. So I think the 30 days is still a thing.

Thanks

[superflexible]

Many thanks for the details, I will investigate what needs to be added to Syncovery to take advantage of this S3 feature fully.

[superflexible]

The new version 7.52 now offers the storage class Infrequent Access, as an alternative to Standard and Reduced Redundancy.