Next, let's check out entries from osd.8's log. Since Ceph OSDs are more numerous than MONs and process a larger rate of transactions we find that they grow much more quickly. Be sure to allocate ample space for them. As above, this cluster is running Ceph's Jewel release.
2017-09-18 00:19:08.268212 7fe07b990700 1 leveldb: Compacting 4@0 + 5@1 files 2017-09-18 00:19:08.327958 7fe07b990700 1 leveldb: Generated table #14159: 55666 keys, 2129686 bytes 2017-09-18 00:19:08.382062 7fe07b990700 1 leveldb: Generated table #14160: 52973 keys, 2129584 bytes 2017-09-18 00:19:08.434451 7fe07b990700 1 leveldb: Generated table #14161: 54404 keys, 2129467 bytes 2017-09-18 00:19:08.488200 7fe07b990700 1 leveldb: Generated table #14162: 54929 keys, ...