Temporarily turn off QoS cgroups on Fedora CoreOS controllers

* Kubelets can hit the ContainerManager Delegation issue and fail
to start (noted in 72c94f1c6). Its unclear why this occurs only
to some Kubelets (possibly an ordering concern)
* QoS cgroups remain a goal
* When a controller node is affected, bootstrapping fails, which
makes other development harder. Temporarily disable QoS on
controllers only. This should safeguard bring-up and hopefully
still allow the issue to occur on some workers for debugging
This commit is contained in:
Dalton Hubble 2019-07-18 23:57:53 -07:00
parent 6cd3e65267
commit 339e323491
2 changed files with 4 additions and 4 deletions

View File

@ -85,8 +85,8 @@ systemd:
--authentication-token-webhook \ --authentication-token-webhook \
--authorization-mode=Webhook \ --authorization-mode=Webhook \
--cgroup-driver=systemd \ --cgroup-driver=systemd \
--cgroups-per-qos=true \ --cgroups-per-qos=false \
--enforce-node-allocatable=pods \ --enforce-node-allocatable="" \
--client-ca-file=/etc/kubernetes/ca.crt \ --client-ca-file=/etc/kubernetes/ca.crt \
--cluster_dns=${cluster_dns_service_ip} \ --cluster_dns=${cluster_dns_service_ip} \
--cluster_domain=${cluster_domain_suffix} \ --cluster_domain=${cluster_domain_suffix} \

View File

@ -86,8 +86,8 @@ systemd:
--authentication-token-webhook \ --authentication-token-webhook \
--authorization-mode=Webhook \ --authorization-mode=Webhook \
--cgroup-driver=systemd \ --cgroup-driver=systemd \
--cgroups-per-qos=true \ --cgroups-per-qos=false \
--enforce-node-allocatable=pods \ --enforce-node-allocatable="" \
--client-ca-file=/etc/kubernetes/ca.crt \ --client-ca-file=/etc/kubernetes/ca.crt \
--cluster_dns=${cluster_dns_service_ip} \ --cluster_dns=${cluster_dns_service_ip} \
--cluster_domain=${cluster_domain_suffix} \ --cluster_domain=${cluster_domain_suffix} \