If user enable fiemap feature in osd side, there exists three ways to make read/write sprase: 1. normal sparse read/write op 2. clone op 3. recover op If any op doesn't support sparse way, it will destroy the advantage of this way and write zero hole to sparse file. Now clone op can support sparse write, this commit(https://github.com/ceph/ceph/pull/2137) implement sparse read/write for recover op. The main advantage is if rbd usage, it will hugely reduce the actual capacity usage and will make clone/recovery faster! On Fri, Sep 26, 2014 at 3:53 PM, Stefan Priebe - Profihost AG <s.priebe at profihost.ag> wrote: > Am 26.09.2014 um 09:36 schrieb Haomai Wang: >> Sorry, the bug last mail reported exists in centos 6.5. >> >> I'm not sure which kernel version solve this bug and 3.12 kernel is >> safe with fiemap on XFS > > What's the advantage of fiemap? I've a vanilla 3.10.55 kernel. > > Stefan > >> On Fri, Sep 26, 2014 at 3:33 PM, Haomai Wang <haomaiwang at gmail.com> wrote: >>> On Fri, Sep 26, 2014 at 3:04 PM, Stefan Priebe - Profihost AG >>> <s.priebe at profihost.ag> wrote: >>>> Hi, >>>> >>>> as the documentation is very unspecific on this option i'm not sure >>>> whether we should enable or disable this one on dumpling. >>> >>> It all depends on your kernel version. For example, as I remember >>> there exists known XFS fiemap bug which will return false extent map. >>> >>> So you must ensure your kernel version with your local filesystem has >>> the right fiemap impl. >>> >>>> >>>> Greets, >>>> Stefan >>>> _______________________________________________ >>>> ceph-users mailing list >>>> ceph-users at lists.ceph.com >>>> http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com >>> >>> >>> >>> -- >>> Best Regards, >>> >>> Wheat >> >> >> -- Best Regards, Wheat
 |