On Tue, Jan 26, 2016 at 5:58 PM, Burkhard Linke <Burkhard.Linke@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote: > Hi, > > On 01/26/2016 10:24 AM, Yan, Zheng wrote: >> >> On Tue, Jan 26, 2016 at 3:16 PM, Burkhard Linke >> <Burkhard.Linke@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> wrote: >>> >>> Hi, >>> >>> On 01/26/2016 07:58 AM, Yan, Zheng wrote: >>> >>> *snipsnap* >>> >>>> I have a few questions >>>> >>>> Which version of ceph are you using? When was the filesystem created? >>>> Did you manually delete 10002af7f78.00000000 from pool 8? >>> >>> >>> Ceph version is 0.94.5. The filesystem was created about 1.5 years ago >>> using >>> a Firefly release. The exact order of commands was: >> >> The FS was created 1.5 years ago can explain the issue. >> >> >> here is way to fix other potential bad files in the FS. >> >> 1. List all object whose name end with .00000000 in pool 7 >> 2. Find all non-directory inodes in the FS >> 3. For all non-directory inodes without corresponding object in pool >> 7, create an empty object in pool 7 and copy xattr 'parent' from >> corresponding object in pool 12 or pool 20. > > Thanks for the help with this case. Most files on that filesystem are > processed by the nightly backup, so other affected files would have shown up > before. > > There're currently abount 10 million objects in pool 7 with over 13 TB > accumulated data, and 15 million objects / 35 TB data in pool 12. Manually > checking them has to be postponed until our cluster is updated with more > suitable hardware. > > I'm going to remove the dangling directory entry from the directory's omap > following the guideline in your first mail if you do not need any further > debug data. another option is create object 10002af7f78.00000000 in pool 7, and copy xattr 'parent' from pool 12 . > > Regards, > Burkhard > > _______________________________________________ > ceph-users mailing list > ceph-users@xxxxxxxxxxxxxx > http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |