SUSE / caasp-salt

A collection of salt states used to provision a kubernetes cluster
Apache License 2.0
64 stars 29 forks source link

[3.0] Uncordon node in a explicit sls action #585

Closed ereslibre closed 6 years ago

ereslibre commented 6 years ago

This way we don't try to uncordon the node in the kubelet/init.sls file, required for example by haproxy, that will end up in the machine trying to early uncordon itself (when haproxy configuration hasn't been written yet, and leading to early failure).

Splitting this action and called only when required (this is: the update process) is safer.

Fixes: bsc#1080978

(cherry picked from commit 02f063385e3a8cd435a76280ca246a87099a01d5)

Backport of https://github.com/kubic-project/salt/pull/584

ereslibre commented 6 years ago

Upgrade is working.

ereslibre commented 6 years ago

Confirmed again with this exact PR: Upgraded a 3+2 cluster directly from 2.0 GM to 3.0:

Summary for admin.infra.caasp.local_master
-------------
Succeeded: 62 (changed=44)
Failed:     0
-------------
Total states run:     62
Total run time: 1109.206 s

and:

admin:~ # kubectl get nodes -o wide
NAME       STATUS    ROLES     AGE       VERSION   EXTERNAL-IP   OS-IMAGE                 KERNEL-VERSION          CONTAINER-RUNTIME
master-0   Ready     master    15m       v1.9.7    <none>        SUSE CaaS Platform 3.0   4.4.126-94.22-default   docker://17.9.1
master-1   Ready     master    15m       v1.9.7    <none>        SUSE CaaS Platform 3.0   4.4.126-94.22-default   docker://17.9.1
master-2   Ready     master    15m       v1.9.7    <none>        SUSE CaaS Platform 3.0   4.4.126-94.22-default   docker://17.9.1
worker-0   Ready     <none>    15m       v1.9.7    <none>        SUSE CaaS Platform 3.0   4.4.126-94.22-default   docker://17.9.1
worker-1   Ready     <none>    15m       v1.9.7    <none>        SUSE CaaS Platform 3.0   4.4.126-94.22-default   docker://17.9.1