Fully persistent allocation info is finally done
Added 2019-04-20 03:39:45 +0000 UTCFinally! It was a huge effort, but it's done and pushed out.
This means that when mounting a filesystem - even after an unclean shutdown - we don't have to walk all the metadata anymore, because it's always updated in a transactional manner and kept fully consistent in the b-tree.
There may be a performance regression for now on multithreaded write workloads, due to lock contention on the alloc btree. But, that will go away when I implement the new btree key cache code (it'll generalize the current deferred btree updates code, used for inode updates).
If anyone does notice a performance regression, post about it so I'll know to prioritize that.
And, this is the last checkbox - besides catching up on the bug reports - on my list before pushing bcachefs upstream. So, expect to see movement on that front over the coming months.
Comments
@Jonathan <a href="https://github.com/koverstreet/bcachefs" rel="nofollow noopener" target="_blank">https://github.com/koverstreet/bcachefs</a>
Tezq Gelrock
2019-05-04 18:39:25 +0000 UTCWow, good news. Now, I'd like to share a idea with you. Add more patron tiers ! Each tier should reflect a certain feature on a wish list. You could use the # of patreon to see what are popular demands. Use multi tier per features ex: 5$-20$-50$-100$-500$ or how you like. PS : I'd like to see native support for SMR 's drives zones as I have one Raidz2 7 drive array of those that i'd like to move to bcachefs. Also while on it, why not move to github, I get slow download speed from your site.
2019-04-27 16:21:08 +0000 UTCThat's terrific news ! Hopefully bcachefs will get a bigger userbase, and perhaps a bigger developper base once it gets upstreamed.
2019-04-25 08:49:49 +0000 UTCCongrats! Can't wait for the upstreaming to happen!
2019-04-21 09:20:07 +0000 UTCThis is amazing, thank you so much!
Lurf Jurv
2019-04-20 18:33:11 +0000 UTC
