We can see the ODF operator's default resource consumption in the source code on GitHub. In the tables described in this document, R refers to requests and L refers to limits. For further information about resource management in Kubernetes, please refer to the upstream documentation.
Since ODF 4.15, different performance profiles have been provided to enhance performance in the ODF cluster. Further information can be found in the Red Hat official documentation.
These are the core pods we have when running ODF in internal mode and their resource consumption:
| Component | Name | OCP object | Number of pods | CPU | Memory |
|---|---|---|---|---|---|
| Operators | ocs-operator | deployment | 1 | R:- / L:- | R:- / L:- |
| Operators | odf-console | deployment | 1 | R:100m / L:100m | R:512Mi / L:512Mi |
| Operators | odf-operator-controller-manager | deployment | 1 | R:200m / L:200m | R:200Mi / L:300Mi |
| Operators | rook-ceph-operator | deployment | 1 | R:- / L:- | R:- / L:- |
| Operators | noobaa-operator | deployment | 1 | R:- / L:250m | R:- / L:512Mi |
| Helpers | ux-backend-server | deployment | 1 | R:- / L:- | R:- / L:- |
| Helpers | ocs-metrics-exporter | deployment | 1 | R:- / L:- | R:- / L:- |
| CSI | csi-rbdplugin | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
| CSI | csi-rbdplugin-provisioner | deployment | 2 | R:- / L:- | R:- / L:- |
| CSI | csi-cephfsplugin | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
| CSI | csi-cephfsplugin-provisioner | deployment | 2 | R:- / L:- | R:- / L:- |
| CSI | csi-addons-controller-manager | deployment | 1 | R:10m / L:1 | R:64Mi / L:512Mi |
| NooBaa | noobaa-endpoint | HPA | Min: 1 - Max: 2 | R:999m / L:999m | R:2Gi / L:2Gi |
| NooBaa | noobaa-core | statefulset | 1 | R:999m / L:999m | R:4Gi / L:4Gi |
| NooBaa | noobaa-db-pg | statefulset | 1 | R:500m / L:500m | R:4Gi / L:4Gi |
| NooBaa | noobaa-backing-store | pod | As many as backing stores | R:100m / L:100m | R:400Mi / L:400Mi |
| Component | Name | OCP object | Number of pods | CPU | Memory |
|---|---|---|---|---|---|
| NFS | nfs | deployment | 1 | R:3 / L:3 | R:8Gi / L:8Gi |
| CSI | csi-nfsplugin | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
| CSI | csi-nfsplugin-provisioner | deployment | 2 | R:- / L:- | R:- / L:- |
| Ceph | rbd-mirror | deployment | 1 | R:1 / L:1 | R:2Gi / L:2Gi |
| Tools | rook-ceph-tools | deployment | 1 | R:- / L:- | R:- / L:- |
These are the core pods we have when running ODF in internal mode using the lean performance profile and their resource consumption:
| Component | Name | OCP object | Number of pods | CPU | Memory |
|---|---|---|---|---|---|
| Ceph | rook-ceph-mgr | deployment | 2 | R:0.5 / L:1 | R:1Gi / L:2Gi |
| Ceph | rook-ceph-mon | deployment | 3 | R:0.5 / L:0.5 | R:1Gi / L:1Gi |
| Ceph | rook-ceph-osd | deployment | As many as devices | R:1.5 / L:1.5 | R:3Gi / L:3Gi |
| Ceph | rook-ceph-mds | deployment | 2 | R:1 / L:1 | R:2Gi / L:2Gi |
| Ceph | rook-ceph-rgw | deployment | 1 | R:1 / L:1 | R:1Gi / L:1Gi |
| Crash | rook-ceph-crashcollector | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
| Exporter | rook-ceph-exporter | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
These are the core pods we have when running ODF in internal mode using the balanced performance profile and their resource consumption:
| Component | Name | OCP object | Number of pods | CPU | Memory |
|---|---|---|---|---|---|
| Ceph | rook-ceph-mgr | deployment | 2 | R:1 / L:2 | R:1.5Gi / L:3Gi |
| Ceph | rook-ceph-mon | deployment | 3 | R:1 / L:1 | R:2Gi / L:2Gi |
| Ceph | rook-ceph-osd | deployment | As many as devices | R:2 / L:2 | R:5Gi / L:5Gi |
| Ceph | rook-ceph-mds | deployment | 2 | R:2 / L:2 | R:6Gi / L:6Gi |
| Ceph | rook-ceph-rgw | deployment | 1 | R:2 / L:2 | R:2Gi / L:2Gi |
| Crash | rook-ceph-crashcollector | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
| Exporter | rook-ceph-exporter | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
These are the core pods we have when running ODF in internal mode using the performance performance profile and their resource consumption:
| Component | Name | OCP object | Number of pods | CPU | Memory |
|---|---|---|---|---|---|
| Ceph | rook-ceph-mgr | deployment | 2 | R:1.5 / L:3 | R:2Gi / L:4Gi |
| Ceph | rook-ceph-mon | deployment | 3 | R:1.5 / L:1.5 | R:3Gi / L:3Gi |
| Ceph | rook-ceph-osd | deployment | As many as devices | R:4 / L:4 | R:8Gi / L:8Gi |
| Ceph | rook-ceph-mds | deployment | 2 | R:3 / L:3 | R:8Gi / L:8Gi |
| Ceph | rook-ceph-rgw | deployment | 1 | R:2 / L:2 | R:4Gi / L:4Gi |
| Crash | rook-ceph-crashcollector | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
| Exporter | rook-ceph-exporter | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
These are the core pods we have when running ODF in external mode and their resource consumption:
| Component | Name | OCP object | Number of pods | CPU | Memory |
|---|---|---|---|---|---|
| Operators | ocs-operator | deployment | 1 | R:- / L:- | R:- / L:- |
| Operators | odf-console | deployment | 1 | R:100m / L:100m | R:512Mi / L:512Mi |
| Operators | odf-operator-controller-manager | deployment | 1 | R:200m / L:200m | R:200Mi / L:300Mi |
| Operators | rook-ceph-operator | deployment | 1 | R:- / L:- | R:- / L:- |
| Operators | noobaa-operator | deployment | 1 | R:- / L:250m | R:- / L:512Mi |
| Helpers | ux-backend-server | deployment | 1 | R:- / L:- | R:- / L:- |
| Helpers | ocs-metrics-exporter | deployment | 1 | R:- / L:- | R:- / L:- |
| CSI | csi-rbdplugin | daemonset | As many as worker nodes | R:- / L:- | R:- / L:- |
| CSI | csi-rbdplugin-provisioner | deployment | 2 | R:- / L:- | R:- / L:- |
| CSI | csi-addons-controller-manager | deployment | 1 | R:10m / L:1 | R:64Mi / L:512Mi |
| NooBaa | noobaa-endpoint | HPA | Min: 1 - Max: 2 | R:999m / L:999m | R:2Gi / L:2Gi |
| NooBaa | noobaa-core | statefulset | 1 | R:999m / L:999m | R:4Gi / L:4Gi |
| NooBaa | noobaa-db-pg | statefulset | 1 | R:500m / L:500m | R:4Gi / L:4Gi |
| NooBaa | noobaa-backing-store | pod | As many as backing stores | R:100m / L:100m | R:400Mi / L:400Mi |