[GH-ISSUE #139] Glacier Archiving Support #84

New issue

Closed

opened 2026-03-04 01:41:52 +03:00 by kerem · 4 comments

kerem commented

2026-03-04 01:41:52 +03:00

Owner

Originally created by @sfrique on GitHub (Mar 5, 2015).
Original GitHub issue: https://github.com/s3fs-fuse/s3fs-fuse/issues/139

Hello,

I would like to save my backups on amazonglacier and i am looking into a easy way to do so.

I think one easy way would be mount a s3 with s3fs and them copy to this mount point.

And them use this s3 feature:
https://aws.amazon.com/pt/blogs/aws/archive-s3-to-glacier/

I want to know if s3fs will have any problem when using this feature of S3

Thanks

Originally created by @sfrique on GitHub (Mar 5, 2015). Original GitHub issue: https://github.com/s3fs-fuse/s3fs-fuse/issues/139 Hello, I would like to save my backups on amazonglacier and i am looking into a easy way to do so. I think one easy way would be mount a s3 with s3fs and them copy to this mount point. And them use this s3 feature: https://aws.amazon.com/pt/blogs/aws/archive-s3-to-glacier/ I want to know if s3fs will have any problem when using this feature of S3 Thanks

kerem closed this issue

2026-03-04 01:41:52 +03:00

kerem commented

2026-03-04 01:41:53 +03:00

Author

Owner

@ggtakec commented on GitHub (Mar 21, 2015):

@sfrique
The setting of the Glacier is based on the assumption that you do from the AWS console without using the REST API, I think s3fs works without any problems except for one issue.
One thing we have to worry about is the stat cache.
I think maybe no problem, but if s3fs does not notice that move the file is moved to the Glacier, you should probably set the stat cache option to off.
Since I have not used the Glacier via s3fs yet, I don't know the result of it, but we expect to be obtained good results.

Regards,

@ggtakec commented on GitHub (Mar 21, 2015): @sfrique The setting of the Glacier is based on the assumption that you do from the AWS console without using the REST API, I think s3fs works without any problems except for one issue. One thing we have to worry about is the stat cache. I think maybe no problem, but if s3fs does not notice that move the file is moved to the Glacier, you should probably set the stat cache option to off. Since I have not used the Glacier via s3fs yet, I don't know the result of it, but we expect to be obtained good results. Regards,

kerem commented

2026-03-04 01:41:53 +03:00

Author

Owner

@sqlbot commented on GitHub (Mar 23, 2015):

@ggtakec it will work until s3fs tries to read objects that have been already been migrated to the glacier storage class. When this migration happens (based on lifecycle policies), the objects remain in the S3 objects listing, since they are still S3 objects (so s3fs should still see them). S3 will return 403 Forbidden with an error class of InvalidObjectState if you try to GET such an object, but HEAD works as expected. Before any GET is possible on such an object, a restore has to be initiated via the API (or console) and this is a limitation users will need to understand, since s3fs should not put itself in a role to do object restorations from glacier, for a number of reasons, not the least of which is the potential for large unexpected charges.

@sqlbot commented on GitHub (Mar 23, 2015): @ggtakec it will work until s3fs tries to read objects that have been already been migrated to the glacier storage class. When this migration happens (based on lifecycle policies), the objects remain in the S3 objects listing, since they are still S3 objects (so s3fs should still see them). S3 will return 403 Forbidden with an error class of InvalidObjectState if you try to GET such an object, but HEAD works as expected. Before any GET is possible on such an object, a restore has to be initiated via the API (or console) and this is a limitation users will need to understand, since s3fs should not put itself in a role to do object restorations from glacier, for a number of reasons, not the least of which is the potential for large unexpected charges.

kerem commented

2026-03-04 01:41:53 +03:00

Author

Owner

@sfrique commented on GitHub (Mar 23, 2015):

Well, if they can be seen, i guess it wil be ok.

The question is, after the restore initiated (and passed the 4 to 5 hours period) the object will be seen by s3fs?

But i don't think i will use s3fs for this purpose anymore.

Thanks anyway!

@sfrique commented on GitHub (Mar 23, 2015): Well, if they can be seen, i guess it wil be ok. The question is, after the restore initiated (and passed the 4 to 5 hours period) the object will be seen by s3fs? But i don't think i will use s3fs for this purpose anymore. Thanks anyway!

kerem commented

2026-03-04 01:41:53 +03:00

Author

Owner

@ggtakec commented on GitHub (Apr 12, 2015):

@sqlbot I can understand about s3fs can get the object on glacier in list by HEAD request.
thanks.

@sfrique I think s3fs can find the object after restoring the object.
But s3fs does not have the function for restoring from glacier.

I closed this issue, and if you have more issue, please reopen this or post new issue.
Thanks in advance for your help.

@ggtakec commented on GitHub (Apr 12, 2015): @sqlbot I can understand about s3fs can get the object on glacier in list by HEAD request. thanks. @sfrique I think s3fs can find the object after restoring the object. But s3fs does not have the function for restoring from glacier. I closed this issue, and if you have more issue, please reopen this or post new issue. Thanks in advance for your help.