You can turn on/off day1 configuration in day1 section, following are the fallback values if day1 is not present.
day1:
workload_partition: true
kdump:
enabled: true
#set blacklist_ice as true on HPE servers
blacklist_ice: false
boot_accelerate: true
#ztp_hub=true will enable rhacm/lvm/gitops/talm on the cluster
ztp_hub: false
#4.12 will ignore this setting, 4.13+ will enable crun by default
crun: true
#4.14, https://issues.redhat.com/browse/OCPBUGS-17660
rcu_normal: true
#4.14, reduce a node boot and fix a race condition issue in sriov operator
sriov_kernel: true
#4.14, sync the node time from ntp when node reboot and ptp got involved
sync_time_once: true
#whether enable cgroup v1; 4.14-4.15: will enable cgv1 by default, 4.16+ will enable cgv2 by default
cgv1: true
container_storage:
enabled: false
device: /dev/nvme0n1
startMiB: 250000
sizeMiB: 0
operators:
ptp:
enabled: true
#if you want to stay on a particular version ptp-operator.4.12.0-202402081808
#version: ptp-operator.4.12.0-202402081808
sriov:
enabled: true
#if you want to stay on a particular version sriov-network-operator.v4.12.0-202402081808
#version: sriov-network-operator.v4.12.0-202402081808
local-storage:
enabled: true
#if you want to stay on a particular version local-storage-operator.v4.12.0-202403082008
#version: local-storage-operator.v4.12.0-202403082008
#preparation work for local storage
provision:
#Get the ID with command: udevadm info -q property --property=ID_PATH /dev/nvme1n1
disk_by_path: pci-0000:c3:00.0-nvme-1
lvs:
1g: 10
2g: 10
4g: 5
5g: 5
10g: 2
15g: 1
30g: 1
gitops:
enabled: false
rhacm:
enabled: false
talm:
enabled: false
mce:
enabled: false
mcgh:
enabled: false
lvm:
enabled: false
fec:
enabled: false
#if you want to stay on a particular version sriov-fec.v2.7.2
#version: sriov-fec.v2.7.2
cluster-logging:
enabled: false
#if you want to stay on a particular version
#version: cluster-logging.v5.8.3
adp:
enabled: false
#set the channel
#channel: stable-1.3
#set the version
#version: 1.3.1
lca:
enabled: false
#set the channel
#channel: stable
#set the version
#version: v4.16.0-89
kubevirt-hyperconverged:
enabled: false
#channel: stable
#set the version
#version: v4.16.3
extra_manifests:
- ${HOME}/1
- ${HOME}/2
- $OCP_Y_VERSIONYou can turn on/off day2 configuration in day2 section.
day2:
# pause the MCP for a certain period and unpause it in order to reduce the number of node reboot.
delay_mcp_update: 60
performance_profile:
enabled: true
#in case you want to specify the performance profile name
name: sno-perfprofile
#optional, default value: true
real_time: false
#optional, present if want to set user_level_networking as true
net:
user_level_networking: true
#optional for hardware tuning, OCP 4.16
hardwareTuning:
isolatedCpuFreq: 2500000
reservedCpuFreq: 2800000
#optional
hugepage:
default: 2M
pages:
- size: 2M
count: 32768
#node: 1
tuned_profile:
enabled: true
#for wrong bios settings, if passive mode is used, set intel_pstate=active
cmdline_pstate: intel_pstate=active
#in case you want to generate kdump for some special scenarios (used in lab)
kdump: false
sysfs:
#cap the intel_pstate peak frequency at 2.5Ghz, used in 4.14. 4.16 can set day2.performance_profile.hardwareTuning
cpufreq_max_freq: 2500000
ptp:
#ptpconfig type: choose any of them: disabled|ordinary|boundary
#chronyd service will be disabled if ordinary or boundary being selected
ptpconfig: disabled
clock_threshold_tuning:
hold_over_timeout: 5
max_offset: 500
min_offset: -500
ordinary_clock:
#name: crdu-ptp-ordinary-clock
interface: ens1f0
boundary_clock:
# only supported for 4.16+, when not enabled, the profiles[0] is used for system clock, 4.15 or earlier versions must set it as false
ha_enabled: true
#name (default): crdu-boundary-clock-ptp-config
#name: crdu-boundary-clock-ptp-config
profiles:
- name: bc-profile-nic1
slave: ens1f0
masters:
- ens1f1
- ens1f2
- ens1f3
- name: bc-profile-nic2
slave: ens2f0
masters:
- ens2f1
- ens2f2
- ens2f3
#enable the cluster monitoring tuning
cluster_monitor_tuning: true
#enable the operator hub tuning: disable unused catalog sources
operator_hub_tuning: true
#disable the network diagnostics
disable_network_diagnostics: true
#4.14 disable the olm pprof(collect-profile cronjob)
disable_olm_pprof: true
#disable the operator auto-upgrade
disable_operator_auto_upgrade: true
#https://github.com/openshift-kni/cnf-features-deploy/blob/master/ztp/source-crs/SriovOperatorConfig-SetSelector.yaml
sriov:
enable_injector: false
enable_webhook: false
lvm:
device_classes:
- name: "sno"
thin_pool_name: "sno-sdb-pool"
selector:
paths:
- "/dev/sdb"
#LocalVolume settings for local storage operator,
#if not present, and if day1.operators.local-storage.provision.lvs is present, will use default name 'local-disks'
# and storageClassName 'general'
local_storage:
local_volume:
name: local-disks
storageClassName: general
extra_manifests:
- ${HOME}/1
- ${HOME}/2
- $OCP_Y_VERSIONAn example to use this repo to create a SNO running on KVM, and install the ZTP required opertators so to act as a ZTP hub.
## Create VM
ssh 192.168.58.14 kcli stop vm hub
ssh 192.168.58.14 kcli delete vm hub -y
ssh 192.168.58.14 'kcli create vm -P uuid=11111111-1111-1111-1234-000000000000 -P start=False -P memory=20480 -P numcpus=16 -P disks=[150] -P nets=["{\"name\":\"br-vlan58\",\"nic\":\"eth0\",\"mac\":\"de:ad:be:ff:10:85\"}"] hub'
ssh 192.168.58.14 kcli list vm
systemctl restart sushy-tools.service
## Generate ISO, config-hub.yaml turn on all ZTP required operators but turn off others
rm -f ~/.cache/agent/image_cache/coreos-x86_64.iso
rm -rf hub
./sno-iso.sh samples/ztp-hub-on-sno/config-hub.yaml
cp hub/agent.x86_64.iso /var/www/html/iso/hub.iso
## Install OCP
./sno-install.sh samples/ztp-hub-on-sno/config-hub.yaml
oc get node --kubeconfig hub/auth/kubeconfig
oc get clusterversion --kubeconfig hub/auth/kubeconfig
echo "Installation in progress, please check it in 30m."
## Run extra manifests
oc --kubeconfig hub/auth/kubeconfig apply -k ./extra-manifests