The State Of The Tux3 File-System For Linux

[phoronix]

Phoronix: The State Of The Tux3 File-System For Linux

Is the Tux3 file-system alive and well for Linux or will it face a fate like Reiser4 where it may never see a mainline state?..

http://www.phoronix.com/vr.php?view=MTMzMjU

[schmidtbag]

I don't get it, what does this filesystem do that other's don't/can't?

[danielbot]

I don't get it, what does this filesystem do that other's don't/can't?

That's a deep question which has been extensively addressed in various posts. See here:

http://phunq.net/pipermail/tux3/

If you are looking for one quick sound bite, hmm, that is hard. Let's just say that we Tux3 developers are bigger fans of the Ext* evolution than various popular next gen efforts both inside and outside the Linux community. But the one killer feature that every true next generation filesystem needs is not even on Ext4's roadmap at the present time: snapshots, and with that replication. This is very definitely on Tux3's roadmap. Plus we have a shiny new toy or two to bring to the party. For one, Tux3's stronger consistency model coupled with higher commit performance seems almost too good to be true, but we have solid measurements to back it up. That by itself is enough excuse for us to keep on pushing forward.

Regards,

Daniel

[matobinder]

Thanks for the info Daniel, I was just about to start searching and/or ask the same question myself.

I don't end up following filesystems stuff as much as I don't do IS work on that level much anymore. When I did, I really liked the AdvFS filesystem that DEC/Compaq/HP had. Now I generally just a user, and our Hitachi NAS and NetApp storage are more like a black box to me.

Having snapshot is a great feature. I'd offer one suggestion on that that I know has bitten us before with the NetApp snapshot. When a volume gets really full, we will put out and message "hey you! delete some stuff you don't need we are at 90%!" That can backfire at times as deleting a ton of data can push snapshot to use much more disk space and we run out even faster. I'm sure there are ways to deal with this that we may not be doing. But it has been my only beef with snapshot.