Current deployment has Sasha/DropZone in support Pod.
StatefulSet gives capibility to dynamically to create EBS volume to attache to newly added Pod/Container. It uses volumeClaimTemplate and StorageClass to achieve this. Delete VolumeClaim will automatically remove the volumes from EC2 otherwise user is responible to clean volumes in EC2.
Compare EBS with EFS, EBS supports to have better performance but when cluster deployed cross multiple AZs it may be difficult to re-use these volumes.
bin/bootstrap-aws.sh"bootstrap-aws.sh" includes apply Kubernetes API RBAC roles and set configmap "hpcc-config" for environment.xml generation. If configmap "hpcc-config" not set user can provide configmap under ./hpcc-config and apply it in ./start script. See aws/configmap/README.md for more detail about configmap settings
./startTo make sure they are up:
kubectl get pods
NAME READY STATUS RESTARTS AGE
dali 1/1 Running 0 100s
efs-provisioner-57965c4946-7w4b5 1/1 Running 0 2d14h
esp-esp1-69b59769bd-hmzss 1/1 Running 0 98s
hpcc-admin 1/1 Running 0 101s
roxie-roxie1-0 1/1 Running 0 97s
roxie-roxie1-1 1/1 Running 0 68s
roxie-roxie2-0 1/1 Running 0 96s
support-0 1/1 Running 0 99s
thor-thor1-0 1/1 Running 0 94s
thor-thor1-1 1/1 Running 0 74s
thormaster-thor1 1/1 Running 0 95sThe cluster should be automatically configured and started. To verify the status
bin/cluster_run.sh status
Status of dali:
mydafilesrv ( pid 972 ) is running ...
mydali ( pid 3722 ) is running ...
Status of esp-esp1-69b59769bd-8gdpp:
mydafilesrv ( pid 990 ) is running ...
esp1 ( pid 5258 ) is running ...
Status of roxie-roxie1-0:
mydafilesrv ( pid 980 ) is running ...
roxie1 ( pid 4154 ) is running ...
Status of roxie-roxie1-1:
mydafilesrv ( pid 978 ) is running ...
roxie1 ( pid 4153 ) is running ...
Status of roxie-roxie2-0:
mydafilesrv ( pid 982 ) is running ...
roxie2 ( pid 4154 ) is running ...
Status of support-0:
mydafilesrv ( pid 1009 ) is running ...
mydfuserver ( pid 7232 ) is running ...
myeclagent ( pid 7427 ) is running ...
myeclccserver ( pid 7616 ) is running ...
myeclscheduler ( pid 7820 ) is running ...
mysasha ( pid 8014 ) is running ...
Status of thor-thor1-0:
mydafilesrv ( pid 973 ) is running ...
Status of thor-thor1-1:
mydafilesrv ( pid 972 ) is running ...
Status of thormaster-thor1:
mydafilesrv ( pid 980 ) is running ...
thor1 ( pid 4208 ) is running with 2 slave process(es) ...Get esp public ic:
kubectl get service
NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
ew-esp1 LoadBalancer 10.100.239.98 a88781525d33911e9a3780efce698321-1790757551.us-east-1.elb.amazonaws.com 8010:32534/TCP 78m
kubernetes ClusterIP 10.100.0.1 <none> 443/TCP 3h12m
svc-roxie-roxie1 ClusterIP None <none> <none> 78m
svc-roxie-roxie2 ClusterIP None <none> <none> 78m
svc-support ClusterIP None <none> <none> 78m
svc-thor-thor1 ClusterIP None <none> <none> 78m
ECLWatch URL: http://a88781525d33911e9a3780efce698321-1790757551.us-east-1.elb.amazonaws.com:8010
Original roxie-roxie1 cluster has 2 instances. To increase it to 4 instances:
kubeclt scale --replicas 4 StatefulSet/roxie-roxie1
kubeclt get pods
ming@BCTWANGXI02-WXL:/mnt/c/work/LexisNexis/Docker_Kubernetes/HPCC-Kubernetes$ kubectl get pods
NAME READY STATUS RESTARTS AGE
dali 1/1 Running 0 76m
esp-esp1-69b59769bd-8gdpp 1/1 Running 0 76m
hpcc-admin 1/1 Running 0 76m
roxie-roxie1-0 1/1 Running 0 76m
roxie-roxie1-1 1/1 Running 0 76m
roxie-roxie1-2 1/1 Running 0 16m
roxie-roxie1-3 1/1 Running 0 15m
roxie-roxie2-0 1/1 Running 0 76m
support-0 1/1 Running 0 76m
thor-thor1-0 1/1 Running 0 76m
thor-thor1-1 1/1 Running 0 76m
thormaster-thor1 1/1 Running 0 76m
To scale it back
kubeclt scale --replicas 2 StatefulSet/roxie-roxie1stop
bin/cluster-run stopstart
bin/cluster-run startGet status
bin/cluster-run status
./stopThis does not delete volumes. Either use AWS Client or go to EC2 console to delete them.