This repository was archived by the owner on Mar 26, 2020. It is now read-only.

Description
Observed behavior
I tried deleting 250 PVCs in a sequential way using below script,
for i in {1..250};do kubectl -n gcs delete pvc pvc$i;done
Below are the observations,
*. All the PVCs are deleted without any issues as expected.
*. Failed to delete 143 PVs, status for each PV shows as Released
*. On glustercli, it failed to delete 146 gluster volumes and are in stopped state.
Expected/desired behavior
PVC delete should delete PVC, PV and teardown gluster volumes.
Details on how to reproduce (minimal and precise)
- Create a 3 node GCS setup using vagrant.
- Create 250 PVCs.
- Delete all the PVCs using below script,
for i in {1..250};do kubectl -n gcs delete pvc $i;done
Information about the environment:
- Glusterd2 version used (e.g. v4.1.0 or master): v6.0-dev.114.gitd51f60b
- Operating system used: Centos 7.6
- Glusterd2 compiled from sources, as a package (rpm/deb), or container:
- Using External ETCD: (yes/no, if yes ETCD version): yes; etcd Version: 3.3.8
- If container, which container image:
- Using kubernetes, openshift, or direct install:
- If kubernetes/openshift, is gluster running inside kubernetes/openshift or outside: Kubernetes
Attaching csi and gluster provisioner logs:
gluster-provisioner-logs.txt
csi-provisioner logs.txt