This document explains the steps for nfs Persistent volume. You might want to use another type of PV (Azure File, GlusterFS, Vsphere volume, Portworx volumes, ScaleIO volumes, etc) but make sure that:
-
the volume supports file locking
-
the mount path on all the nodes will be the same
-
you have the libraries or tools installed on the Workers.
For instance, with nfs, make sure that your workers (each of them) will have access to net-tools and nfs-utils:
yum install -y net-tools nfs-utils nfs-utils-lib
Make sure you are on "default" Namespace.
Click on “+Create” then “Create from file” and upload “nfs-volume.yaml” here - make sure that you update the file (replace value starting with @@@@)
Validate that the Persistent Volume Claims is added
Make sure you are on "default" Namepace.
Click on “+Create” then “Create from fileand upload "boomi-molecule.yaml" here - make sure that you update the file (replace value starting with @@@@)
You will see the new deployment with 1 Pod requested:
Refresh the screen
Check the content of the NFS location and the content of hosts.log
nas_server# ll
total 4
drwxrwxrwx 1 boomi users 70 Aug 23 10:06 .
drwxrwxrwx 1 ant users 10 Aug 19 18:28 ..
-rw-r\--r\-- 1 admin users 286 Aug 23 10:06 hosts.log
drwxr-xr-x 1 admin users 122 Aug 23 10:07 Molecule\_centos\_kubernetes
nas_server# cat hosts.log
Fri Aug 23 02:10:17 UTC 2019: Starting molecule-deployment-57885b48f6-8cmpw with IP 10.47.0.1 with user uid=0(root) gid=0(root) groups=0(root)
Connect to AtomSphere and check that you have a new element
Wait for few minutes and you will see the Atom up
If the Atom/Molecule doesn't appear or if bullet remains gray, please check the logs Kubernetes and also the logs contained in the Shared Folder, you should have something similar:
nas_server# cat Molecule_centos_kubernetes/install_Molecule_centos_kubernetes.log
No suitable Java Virtual Machine could be found on your system.
Downloading JRE with wget ...
Unpacking JRE ...
Preparing JRE ...
Starting Installer ...
Warning: Java not found in the system PATH.
It may cause an issue during Atom install.
Please update your PATH environment variable.
Example:
export PATH=/path/to/jre/bin:$PATH
Authenticating credentials
The installation directory has been set to /var/boomi/Molecule_centos_kubernetes.
Retrieving Build Number
Extracting files...
Downloading JRE Files
Downloading Atom Files
Downloading Molecule Files
Retrieving Container ID
Retrieving account keystore.
Retrieving account trustore.
Configuring Molecule.
Finishing installation...
Attach the Molecule to an environment
Check the cluster status
Return to Kubernetes Dashboard and select "Deployments" option:
Select the right menu and "Scale"
Change the number of desired pods to 3
After 1 minute or two, go back AtomSphere and check the Cluster Status
For auto-scaling based on CPU Load (for instance 25%), connect to your Kubernetes master and run the following command:
kubectl autoscale deployment molecule-deployment --cpu-percent=25 --min=1 --max=3
Change the type of API Type to Advanced.
Deploy one of your API (REST or SOAP) on the environment:
Then test your API with your Browser/SOAPUI or Postman using the following URI:
http://localhost:8001/api/v1/namespaces/default/services/http:molecule-service:9090/proxy/<API Resource Path>
Run as many calls as the number of Pods:
View the extended information for each call, and validate that the Node ID is changing for each call:
-
Extended Execution Information for 1st call: Node ID kubeworker3
-
Extended Execution Information for 2nd call: Node ID kubeworker2
-
Extended Execution Information for 3rd call: Node ID kubeworker1