Skip to content
This repository has been archived by the owner on May 23, 2023. It is now read-only.

Multi client support on one instance

Sven Ehlert edited this page Mar 2, 2015 · 7 revisions

Planned feature

One AWS instance should host multiple clients and multiple client implementations

AWS Instance naming: host-{hostnr}, from 00000 - 99999 -> up to 10000 hosts

Each host can host up to 1000 clients of the same implementation

client naming: client-{client_implementation}-{clientnr}

An individual client can then be addressed something like: host-05464-client-go-365.

Port mapping:

  • 30000 + yyy -> p2p port of go client

  • 31000 + yyy -> p2p port of cpp client

  • 32000 + yyy -> p2p port of python client

  • 20000 + yyy -> json-rpc port of go client

  • 21000 + yyy -> json-rpc port of cpp client

  • 22000 + yyy -> json-rpc port of python client

Steps:

  • Ansible playbook addressing multiple clients per hosts
  • setup a process manager to handle docker logs (ansible does support this only for single commands)
  • change logic to generate node id
  • refactor python controller semantic,now a host != a client
  • refactor scenarios to use the new semantic.
Clone this wiki locally