In GCP, the Elastifile cluster is configured in VPC A, and the clients are configured in VPC B.
Both in the same zone and project.
Step by Step Guide
- Configure Internal Load Balancer in VPC A.
Health checks with the following configurations:
- using tcp and port 111
- interval - 5 sec
- timeout - 5 sec
- threshold - 3
- Create an Instance Group in VPC A, with the ECFS nodes. see the following image for example:
- Add the Instance Group (from step 2) to the Internal Load Balancer Backend.
- Add the following ports to the Internal Load Balancer Frontend:
- See the following image for Internal Load Balancer, for example:
Create two VPC Network Peering, one for VPC A --> VPC B, and the second for VPC B --> VPC A. see the following image for example:
Once the Status is Connected, you can mount the clients from VPC B, to the ECFS in VPC A, using the IP address of the ILB.
Any new instance (node) should be added to the instance group created in step 2 and deleted one should be removed.
Note: Ping or Showmount will not work. ( due to current limitation of the GCP ILB supporting either TCP or UDP...)