Using Kubernetes Engine to Deploy Apps with Regional Persistent DisksGo to Lab
In case, the mariadb pvc is created in the same zone as the wordpress pod, the setup can not recover from zone failure, as the mariadb container will be pending with node affinity conflict (missing pvc with standard storage class).
There was an error at the end, after simulating the failure, and the cluster couldn't recover (move the pod and add a node), hence the wordpress app stopped working after the failure. The cluster details show a size of 3 but there was only 2 nodes. The pod restarted 3 times and was showing the status `CrashLoopBackOff`. The version of k8s is `1.14.7-gke.10` and my project id is `qwiklabs-gcp-00-8e3060e0c3e6`
This lab is broken. The mariadb volume is only in one zone, and if that is the zone you kill the instace group in, it is unable to reschedule, and thus even though wordpress was able to move, it won't be able to contact the database.
mariadb was using a standard persistent disk and lost access to its storage...
Doesn'r work word press after deleting the instance group