Re:Deduplicating a 10.4 TiB game preservation archive (WIP)
by hamandcheese on Saturday December 21, 2024 @03:09PM
(#42481931)
Why a content-addressable blob rather than just a directory with entries named after the SHA (which you can later tar up if desired)?
[ Parent
| on HN ]
Re:Deduplicating a 10.4 TiB game preservation archive (WIP)
by rgovostes on Saturday December 21, 2024 @10:22PM
(#42484179)
That would work as well. I'm not sure if extraction of a file by filename from a compressed tarball is particularly efficient or not.
Re:Deduplicating a 10.4 TiB game preservation archive (WIP)
by hamandcheese on Sunday December 22, 2024 @02:03AM
(#42484830)
Ahh I see. If fast random access is a requirement that makes things a bit more tricky, yes.