1 Deploy gfs env in k8s by run cmd gk-deploy -g -n gfs-infra -c kubectl -v
2 Delete gfs env by run cmd gk-deploy -g -n gfs-infra --abort, after cmd exec finished, some gfs pod still in 'Terminating' status:
[root@host-10-1-236-142 deploy]# kubectl get pods -n gfs-infra --watch
NAME READY STATUS RESTARTS AGE
deploy-heketi-5f6c465bb8-49rwx 1/1 Running 0 47s
glusterfs-9bcp4 1/1 Terminating 0 18m
glusterfs-pkvjd 1/1 Terminating 0 18m
3 Deploy gfs env again by run cmd gk-deploy -g -n gfs-infra -c kubectl -v, deploy will fails like below:
......
Determining heketi service URL ... OK
kubectl -n gfs-infra exec -i deploy-heketi-5f6c465bb8-49rwx -- heketi-cli -s http://localhost:8080 --user admin --secret '' topology load --json=/etc/heketi/topology.json 2>&1
Creating cluster ... ID: 3e33eca5d33cfc36a1c3e481c4eacbc3
Allowing file volumes on cluster.
Allowing block volumes on cluster.
Creating node host-10-1-236-142 ... Unable to create node: New Node doesn't have glusterd running
Creating node host-10-1-241-96 ... Unable to create node: New Node doesn't have glusterd running
Creating node host-10-1-241-229 ... Unable to create node: New Node doesn't have glusterd running
.....
That because gk-deploy think gfs pods still exist, but actually they cannot work well.
So, I think gk-deploy need check glusterd pod's Terminating status, or we can wait this process done.
1 Deploy gfs env in k8s by run cmd gk-deploy -g -n gfs-infra -c kubectl -v 2 Delete gfs env by run cmd gk-deploy -g -n gfs-infra --abort, after cmd exec finished, some gfs pod still in 'Terminating' status: [root@host-10-1-236-142 deploy]# kubectl get pods -n gfs-infra --watch NAME READY STATUS RESTARTS AGE deploy-heketi-5f6c465bb8-49rwx 1/1 Running 0 47s glusterfs-9bcp4 1/1 Terminating 0 18m glusterfs-pkvjd 1/1 Terminating 0 18m 3 Deploy gfs env again by run cmd gk-deploy -g -n gfs-infra -c kubectl -v, deploy will fails like below: ...... Determining heketi service URL ... OK kubectl -n gfs-infra exec -i deploy-heketi-5f6c465bb8-49rwx -- heketi-cli -s http://localhost:8080 --user admin --secret '' topology load --json=/etc/heketi/topology.json 2>&1 Creating cluster ... ID: 3e33eca5d33cfc36a1c3e481c4eacbc3 Allowing file volumes on cluster. Allowing block volumes on cluster. Creating node host-10-1-236-142 ... Unable to create node: New Node doesn't have glusterd running Creating node host-10-1-241-96 ... Unable to create node: New Node doesn't have glusterd running Creating node host-10-1-241-229 ... Unable to create node: New Node doesn't have glusterd running ..... That because gk-deploy think gfs pods still exist, but actually they cannot work well. So, I think gk-deploy need check glusterd pod's Terminating status, or we can wait this process done.