Reduce ceph memory overhead for AIO by setting is_hci to true
This reserves 80% of the host memory to *not* be used when calculating memory limits for the ceph osd processes. This setting is used in ceph-ansible for 'hyperconverged' deployments where nova compute is co-located with osd on the same nodes and the majority of memory must be available for vm guests rather than storage. Change-Id: I7bba3429d71b30253bc72fd013cec4d085eb8fbf
This commit is contained in:
parent
df97c1747d
commit
76cd97b1e5
@ -14,6 +14,7 @@
|
|||||||
# limitations under the License.
|
# limitations under the License.
|
||||||
|
|
||||||
## ceph-ansible AIO settings
|
## ceph-ansible AIO settings
|
||||||
|
is_hci: true
|
||||||
common_single_host_mode: true
|
common_single_host_mode: true
|
||||||
monitor_interface: "{{ ('metal' in bootstrap_host_scenarios_expanded) | ternary('br-mgmt', 'eth1') }}" # Management network in the AIO
|
monitor_interface: "{{ ('metal' in bootstrap_host_scenarios_expanded) | ternary('br-mgmt', 'eth1') }}" # Management network in the AIO
|
||||||
public_network: "{{ (mgmt_range ~ '.0/' ~ netmask) | ipaddr('net') }}"
|
public_network: "{{ (mgmt_range ~ '.0/' ~ netmask) | ipaddr('net') }}"
|
||||||
|
Loading…
Reference in New Issue
Block a user