The problem seems similar to https://tracker.ceph.com/issues/23871 which was fixed in mimic but not luminous: fe5038c7f9 osd/PrimaryLogPG: clear data digest on WRITEFULL if skip_data_digest .. dan On Fri, Jul 13, 2018 at 12:45 PM Dan van der Ster <dan@xxxxxxxxxxxxxx> wrote: > > Hi, > > Following the reports on ceph-users about damaged cephfs after > updating to 12.2.6 I spun up a 1 node cluster to try the upgrade. > I started with two OSDs on 12.2.5, wrote some data. > Then I restarted the OSDs one by one while continuing to write to the > cephfs mountpoint. > Then I restarted the (single) MDS, and it is indeed damaged with a crc error: > > 2018-07-13 12:38:55.261379 osd.1 osd.1 137.138.62.86:6805/35320 2 : > cluster [ERR] 2.15 full-object read crc 0xed77af7c != expected > 0x1a1d319d on 2:aa448500:::500.00000000:head > 2018-07-13 12:38:55.285994 osd.0 osd.0 137.138.62.86:6801/34755 2 : > cluster [ERR] 2.13 full-object read crc 0xa73a97ef != expected > 0x3e6fdb4a on 2:c91d4a1d:::mds0_inotable:head > > I think it goes without saying that nobody should upgrade a cephfs to > 12.2.6 until this is understood. > > -- Dan _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |