Holger Hoffstätte posted on Sun, 23 Nov 2014 22:49:01 +0000 as excerpted: > On Sun, 23 Nov 2014 13:16:50 -0800, Marc MERLIN wrote: > > (snip) > >> That makes sense. I'll try to synthetize all this and rewrite my blog >> post and the wiki to make this clearer. > > Maybe also add that as of 3.18 empty block groups are automatically > collected, so balancing to prevent ENOSPC-by-empty-chunks is no longer > necessary. This works pretty well; I haven't run balance in weeks, and > my total-vs.-used overhead has always been <10 GB. For those of us who have been around btrfs for awhile, this still sounds like the stuff if science fiction, perhaps possible sometime in the future, but definitely not something we're yet used to having actually automatically handled for us. =:^) So I think we're all glad it's here now, but kind of holding our breath waiting for that bug due to this that stops everything cold. I think the feeling is, OK, but let's not rock the boat too much in our haste to celebrate, or we might find ourselves unexpectedly in the water once again. Let's just go on teaching people to swim, assuming they'll need to know how, and if they never do, well then that's a bonus! =:^) But I think once we get into the 3.19 development cycle, if there's no critical bugs with this 3.18 feature yet, then and only then are we likely to start really talking about it. =:^) -- Duncan - List replies preferred. No HTML msgs. "Every nonfree program has a lord, a master -- and if you use the program, he is your master." Richard Stallman -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
