openstack/kolla-ansible
Code Issues Proposed changes

optimize qemu.conf

Browse Source 
bump up the max_files to 32768 and max_processes to 131072.
when nova used ceph as backend, the default limit 1024 is not enough.
each connection from rbd image to osd needs 1 fd and 2 threads. if we
have 200 osds, we need 200 fds and 400 threads for 1 image.

Change-Id: I94c3ec111473ea2ccacdea5dbbf3fdc9c569859f
This commit is contained in:
binhong.hua 2019-01-17 15:13:42 +08:00 committed by Lucas.hua
parent 20c58a9c53
commit 023334dfcb
3 changed files with 17 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
ansible/roles/nova/defaults/main.yml
View File

@ -397,6 +397,13 @@ nova_enabled_notification_topics: "{{ nova_notification_topics | selectattr('ena
vmware_vcenter_datastore_regex: ".*"
ovs_bridge: "nsx-managed"
####################
# Libvirt/qemu
####################
# The number of max files qemu can open
qemu_max_files: 32768
# The number of max processes qemu can open
qemu_max_processes: 131072
####################
# Kolla

3
ansible/roles/nova/templates/qemu.conf.j2
View File

@ -2,3 +2,6 @@ stdio_handler = "file"
user = "nova"
group = "nova"
max_files = {{ qemu_max_files }}
max_processes = {{ qemu_max_processes }}

7
releasenotes/notes/add_qume_options-8e9f040d0bc36fe6.yaml Normal file
View File

@ -0,0 +1,7 @@
---
features:
- |
Bump up the max_files to 32768 and max_processes to 131072.
when nova used ceph as backend, the default limit 1024 is not enough.
each connection from rbd image to osd needs 1 fd and 2 threads. if we
have 200 osds, we need 200 fds and 400 threads for 1 image.