How to monitor an OpenNebula datastore

Storage pressure in OpenNebula often appears as image uploads that stall, VM deployments that remain pending, or backup jobs that cannot reserve space. Datastore monitoring ties those symptoms to capacity, state, image count, and the storage drivers that move disks between the front-end, datastores, and hypervisor hosts.

The onedatastore CLI shows both the datastore pool summary and the detailed object record. The list view is the fastest check for STAT, SIZE, AVA, TYPE, DS, and TM, while the show view exposes the base path, capacity block, driver attributes, and stored image IDs for one datastore.

System datastores need a separate check when local storage is involved, because the datastore summary can show no global size while each host has its own local system datastore capacity. OpenNebula refreshes datastore values on the daemon's datastore monitoring interval, so compare OpenNebula output with the relevant host or filesystem view before treating a stale number as current free space.

1. List the datastore pool.

   $ onedatastore list
     ID NAME                SIZE AVA CLUSTERS IMAGES TYPE DS      TM      STAT
      2 files                50G 86% 0             0 fil  fs      local   on
      1 default              50G 86% 0             2 img  fs      local   on
      0 system                 - -   0             0 sys  -       local   on

   STAT should show on for datastores available to normal operations. SIZE and AVA show the monitored capacity for datastores that expose a pool-wide capacity value.

2. Show the target datastore record.

   $ onedatastore show default
   DATASTORE 1 INFORMATION
   ID             : 1
   NAME           : default
   USER           : oneadmin
   GROUP          : oneadmin
   CLUSTERS       : 0
   TYPE           : IMAGE
   DS_MAD         : fs
   TM_MAD         : local
   BASE PATH      : /var/lib/one//datastores/1
   DISK_TYPE      : FILE
   STATE          : READY
   
   DATASTORE CAPACITY
   TOTAL:         : 50G
   FREE:          : 43.2G
   USED:          : 6.8G
   LIMIT:         : -
   
   DATASTORE TEMPLATE
   DS_MAD="fs"
   TM_MAD="local"
   TYPE="IMAGE_DS"
   
   IMAGES
   0
   1

   DS_MAD identifies the datastore driver, and TM_MAD identifies the transfer manager used when VM disks are cloned, linked, moved, or deleted.

3. Check which images are stored in the target image or files datastore.

   $ oneimage list
     ID USER     GROUP    NAME            DATASTORE     SIZE TYPE PER STAT RVMS
      0 oneadmin oneadmin ubuntu-24.04    default       3.1G OS   No rdy     0
      1 oneadmin oneadmin app-data        default        20G DB   Yes rdy     1
      2 oneadmin oneadmin init.iso        files         720M CD   No rdy     0

   The DATASTORE column should match the datastore being checked, and RVMS shows whether an image is referenced by running VMs.

4. List hosts before checking a local system datastore.

   $ onehost list
     ID NAME        CLUSTER     RVM   TCPU   FCPU   ACPU   TMEM   FMEM   AMEM STAT
      0 host01      default       3    800    520    800  31.1G  18.2G  31.1G   on
      1 host02      default       2    800    610    800  31.1G  21.4G  31.1G   on

5. Check local system datastore capacity on the host that may receive the VM.

   $ onehost show host01
   HOST 0 INFORMATION
   ID                    : 0
   NAME                  : host01
   CLUSTER               : default
   STATE                 : MONITORED
   IM_MAD                : kvm
   VM_MAD                : kvm
   LAST MONITORING TIME  : 06/25 09:35:18
   
   HOST SHARES
   RUNNING VMS           : 3
   ##### snipped #####
   
   LOCAL SYSTEM DATASTORE #0 CAPACITY
   TOTAL:                : 467.7G
   USED:                 : 113.5G
   FREE:                 : 354.2G

   Use onehost show when the system datastore reports SIZE and AVA as - or when the TM driver uses host-local storage.

6. Compare a file-backed datastore with the backing filesystem on the server that owns the base path.

   $ df -h /var/lib/one/datastores/1
   Filesystem      Size  Used Avail Use% Mounted on
   /dev/vdb1        50G  6.8G   44G  14% /var/lib/one/datastores

   For Ceph, SAN, backup, or remote object storage drivers, use the storage platform's own capacity tools instead of treating the front-end filesystem as the datastore backend.

7. Recheck the datastore summary after the next datastore monitoring cycle.

   $ onedatastore list
     ID NAME                SIZE AVA CLUSTERS IMAGES TYPE DS      TM      STAT
      2 files                50G 86% 0             0 fil  fs      local   on
      1 default              50G 86% 0             2 img  fs      local   on
      0 system                 - -   0             0 sys  -       local   on

   MONITORING_INTERVAL_DATASTORE in /etc/one/oned.conf controls the datastore monitoring cycle. If the value remains stale after the next cycle, inspect /var/log/one/oned.log and the opennebula service before changing datastore placement or capacity limits.
   Related: How to manage OpenNebula services with systemctl in Linux