Skip to content

Commit f696db5

Browse files
committed
Reduce service workers for aufn-ceph & virtualised ci-multinode envs
When running a multinode environment in VMs, we can easily run out of memory on controllers. Reduce to 2 service/RPC workers in this case. Also reduce RPC workers to 1 for the aufn-ceph environment.
1 parent 658503d commit f696db5

File tree

2 files changed

+6
-0
lines changed

2 files changed

+6
-0
lines changed

etc/kayobe/environments/aufn-ceph/kolla/globals.yml

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -7,6 +7,7 @@ nova_compute_virt_type: qemu
77
# Reduce the control plane's memory footprint by limiting the number of worker
88
# processes to one per-service.
99
openstack_service_workers: "1"
10+
openstack_service_rpc_workers: "1"
1011

1112
glance_backend_ceph: "yes"
1213
cinder_backend_ceph: "yes"

etc/kayobe/environments/ci-multinode/kolla/globals.yml

Lines changed: 5 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,9 @@
11
---
2+
# Reduce the control plane's memory footprint by limiting the number of worker
3+
# processes to two per-service when running in a VM.
4+
openstack_service_workers: "{% raw %}{{ [ansible_facts.processor_vcpus, 2 if ansible_facts.virtualization_role == 'guest' else 5] | min }}{% endraw %}"
5+
openstack_service_rpc_workers: "{% raw %}{{ [ansible_facts.processor_vcpus, 2 if ansible_facts.virtualization_role == 'guest' else 3] | min }}{% endraw %}"
6+
27
# Glance Ceph configuration
38
glance_backend_ceph: "yes"
49

0 commit comments

Comments
 (0)