[GH-ISSUE #1611] [Question]multiple clients mounting the same bucket #847

New issue

Closed

opened 2026-03-04 01:49:19 +03:00 by kerem · 2 comments

kerem commented

2026-03-04 01:49:19 +03:00

Owner

Originally created by @GoTheExtraMile on GitHub (Mar 26, 2021).
Original GitHub issue: https://github.com/s3fs-fuse/s3fs-fuse/issues/1611

Hi !
in limitations your say :

non-AWS providers may have eventual consistency so reads can temporarily yield stale data (AWS offers read-after-write consistency since Dec 2020)
no coordination between multiple clients mounting the same bucket
`
Can I think that when using Amazon storage, it is highly consistent for multiple clients to mount the same bucket?

Originally created by @GoTheExtraMile on GitHub (Mar 26, 2021). Original GitHub issue: https://github.com/s3fs-fuse/s3fs-fuse/issues/1611 Hi ! in limitations your say : ` - non-AWS providers may have eventual consistency so reads can temporarily yield stale data (AWS offers read-after-write consistency since Dec 2020) - no coordination between multiple clients mounting the same bucket ` Can I think that when using Amazon storage, it is highly consistent for multiple clients to mount the same bucket?

kerem closed this issue

2026-03-04 01:49:19 +03:00

kerem commented

2026-03-04 01:49:20 +03:00

Author

Owner

@gaul commented on GitHub (Mar 28, 2021):

s3fs has other sources of inconsistency due to caching. For example, if client1 calls stat on a file then client2 appends to it, client1 will see the old metadata until stat_cache_expire expires, by default 900 seconds. Even if you reduce this to 0 there is no coordination between the s3fs clients. Further s3fs has close-to-open semantics meaning that client1's writes will not be published to client2 until a close or fsync. In other network file systems like SMB2, a client can create leases to files which ensures that other clients cannot access them until the lease expires or is revoked.

@gaul commented on GitHub (Mar 28, 2021): s3fs has other sources of inconsistency due to caching. For example, if client1 calls stat on a file then client2 appends to it, client1 will see the old metadata until `stat_cache_expire` expires, by default 900 seconds. Even if you reduce this to 0 there is no coordination between the s3fs clients. Further s3fs has close-to-open semantics meaning that client1's writes will not be published to client2 until a close or fsync. In other network file systems like SMB2, a client can create leases to files which ensures that other clients cannot access them until the lease expires or is revoked.

kerem commented

2026-03-04 01:49:20 +03:00

Author

Owner

@GoTheExtraMile commented on GitHub (Mar 31, 2021):

thanks

@GoTheExtraMile commented on GitHub (Mar 31, 2021): thanks