The mystical Ceph available space and disk usage

Posted on

Hello,

I am running a ceph cluster for testing purposes and it's looking very nice, except some lack of information that I am unable to clarify...

I am running Mimic - last version - on a 10 servers with 22SSDs of 1T each searver.

I've started adding some data to see the load, IO speeds and all but after a closer look I cannot understand how the space used and free is calculated.

I have only one RBD with replication 2, CephFS replication 2, and another rbd pool with also replication 2.

My question is... how is the space calculated ? I 've added everything and there is something missing and I cannot understand what.

This is how ceph df looks like:

ceph df
GLOBAL:
    SIZE        AVAIL      RAW USED     %RAW USED 
    141 TiB     61 TiB       80 TiB         56.54 
POOLS:
    NAME                           ID     USED        %USED     MAX AVAIL     OBJECTS   
    rbd                            1       23 TiB     51.76        22 TiB       6139492 
    .rgw.root                      7      1.1 KiB         0        22 TiB             4 
    default.rgw.control            8          0 B         0        22 TiB             8 
    default.rgw.meta               9      1.7 KiB         0        22 TiB            10 
    default.rgw.log                10         0 B         0        22 TiB           207 
    default.rgw.buckets.index      11         0 B         0        22 TiB             3 
    default.rgw.buckets.data       12      61 KiB         0        22 TiB            11 
    default.rgw.buckets.non-ec     13         0 B         0        22 TiB             0 
    cephfs_data            14      12 TiB     35.42        22 TiB     296419491 
    cephfs_metadata        15     174 MiB         0        11 TiB       1043395 
    rbd2                     16     2.7 TiB     11.18        22 TiB        720971