On 2019/12/21 下午8:07, Swâmi Petaramesh wrote: > Le 21/12/2019 à 13:00, Qu Wenruo a écrit : >> A known regression introduced in v5.4. >> >> The new metadata over-commit behavior conflicts with an existing check >> in btrfs_statfs(). >> It is completely a runtime false behavior, and had*no* other bad effect. > > Many thanks to you for the quick explanation. > > I understand « no other bad effect » as « The FS will be OK again when > this is fixed in the kernel, I can leave the disk in my drawer in the > mean time » ? You can even use it without any problem with v5.4 kernel, except the `df` will sometimes bother you. > > Well I find that makes several recent problematic regressions in > BTRFS... 5.2... 5.4... That's why we have LTS, and commercial company baked kernels. Thanks, Qu > > I'm starting worrying about this development model. > > Something as serious as an « Enterprise grade filesystem » that has been > under heavy development for about ten years is not supposed to be a toy > that breaks every now and then and couple patches is it ? > > Or it would mean that it would ever stay a toy, not an enterprise-grade > FS... > > Kind regards. >
