We are using LakeFS Server, that provide S3 compatible API and Azure Blob Storage.
We are using rclone mount
to access to those data store on our Linux servers
LakeFS data is immutable (with protected branch): file don't get changed, created, deleted. For our Azure blob storage, because of the way we upload the data, it is also immutable like for Lakefs. If a file changes, that an issue from our upload end, which is out of scope of this topic
I would like to get the best performance of caching system knowing the above. I had read the docs and forum and found those options that I think it can be useful:
--attr-timeout 52w
: no more kernel call to rclone once file is cached once ?
--dir-cache-time 52w
: no more cache invalidation for directories ?
--vfs-fast-fingerprint
: will this remove all fingreprinting ? I should not need fingreprinting right ?
--vfs-cache-mode full
--vfs-cache-max-age 52w
: no more cache eviction ?
--no-checksum --no-modtime
: do I need those with --vfs-fast-fingerprint
?
--read-only
: good to have as the system is immutable anyway ?
--transfers 64
: useful for parallel file read ?
Is there any option that I missed ?
Thank you for your help.