Some osd node can not boot on a host of my cluster, the reason from osd.log maybe different, looks like journal broken.
logs attached, I can attach gziped core dump if needed. BTW, is ceph has a fsck.* like tool to check the filesystem consistency by hand ?
osd.22.log.gz
Description: GNU Zip compressed data
osd.28.log.gz
Description: GNU Zip compressed data
osd.32.log.gz
Description: GNU Zip compressed data
