Test Plan for RNode

Purpose of this Document

Networking Features

The Node starts up in bootstrap node when configured to do so
The node connects to peers when configured in networking mode
The node starts up in standalone mode

Rholang CLI (Interpreter)

Interpreter can be invoked via gRPC API
Rholang Code can be evaluated in the Node regardless of the mode the node boots up in
- standalone mode
- Bootstrap mode
- Connected mode
Rholang CLI can be invoked in 2 modes
- REPL
  - Execute: @"listCh"!([1, 2, 3]) | for(@list <- @"listCh"){ match list { [a, b, c] => { @"stdout"!(a) } } }
- EVAL - accepts a file

STDOUT channel works:

@"stdout"!("foo") returns 'foo' to STDOUT Channel.

gRPC API (Node API)

Server accepts requests and returns responses
List of peers is accurate & matches what is presented on the server console.
- Response from peers reflects peers currently connected and changes as the peer list changes
Deploy-demo call works (--deploy-demo)

Metrics

Check Promethus server
Check that all metrics in the release appear

Installation, Starting and Ending the program

Installation completes without issues
Ctrl-C terminates the program without throwing an exception
Program starts without exceptions (providing that the directories needed exist on the system)
In the event that the bootstrap node cannot be found, system provides a message an ends normally.

Debian Instructions

git clone https://github.com/rchain/rchain.git
cd rchain
git checkout dev
sbt clean bnfc:generate node/debian:packageBin
sudo apt install ./node/target/
sudo apt install ./node/target/rnode_0.3.1_all.deb
sudo systemctl start rnode.service
# in one terminal
sudo journalctl -f -u rnode.service
# in another terminal (for repl testing)
rnode --repl
# after that (for deploy testing)
rnode --deploy-demo

Fedora Instructions

git clone https://github.com/rchain/rchain.git
cd rchain
git checkout dev
sbt clean bnfc:generate node/rpm:packageBin
sudo dnf install node/target/rpm/RPMS/noarch/rnode-0.3.1-1.noarch.rpm
sudo systemctl start rnode.service
# in one terminal
sudo journalctl -f -u rnode.service
# in another terminal (for repl testing)
rnode --repl
# after that (for deploy testing)
rnode --deploy-demo

Clean-up scripts (need a bit of improvement, but should get you started):

https://gist.github.com/henrytill/ac72d8616e52d836efc3008bbd9f67dd

Run this at the REPL as many times as humanly possible:

@"stdout"!("foo")

Run this to evaluate rholange example files

Eval examples

for i in $(ls /usr/share/rnode/examples/*.rho); do echo $i; rnode --eval $i; done

Testing Matrix

Who?	What Test the tasks described above in your assigned environment below.
Jeremy	Ubuntu 16.04 LTS Results (DEB): Installed cleanly Started systemd service succesfully Tested REPL Tested EVAL Tested Deploy Demo Comments:	Fedora 27 VM Results (RPM): Installed cleanly Started systemd service succesfully Tested REPL Tested EVAL Tested Deploy Demo Comments:	Debian Stable VM Results (DEB): Installed cleanly Started systemd service succesfully Tested REPL Tested EVAL Tested Deploy Demo Comments:	I can't see my cursor/text after I exit using CTRL-C from REPL 1GB Ubuntu VM 2GB Debian VM 4GB Fedora VM Ubuntu LTS baremetal 32GB
Henry	Debian Stable VM (1 GB RAM, 1 CPU): (DEB, Binary Tarball) Results (Binary Tarball): Started node manually Tested REPL Tested EVAL Tested Deploy Demo Results (DEB): Installed cleanly Started systemd service succesfully Tested REPL Tested EVAL Tested Deploy Demo	Fedora Laptop (8 GB RAM, 4 CPU) (RPM, Binary Tarball) Results (Binary Tarball): Started node manually Tested REPL Tested EVAL Tested Deploy Demo Results (RPM): Installed cleanly Started systemd service succesfully Tested REPL Tested EVAL Tested Deploy Demo		Comments: Still seeing NPE when I exit REPL with CTRL-C CORE-485 - Getting issue details... STATUS Eval works "as advertised" but it is really counterintuitive to have to provide the absolute path to a file on the server, rather than client. Jeremy and I evaluated every file in the top-level of `/usr/share/rnode/examples` in networked mode and all is good We should exit immediately with status code >= 1 and error message when libsodium is not found: CORE-516 - Getting issue details... STATUS
Medha	Test docker 1 CPU, 1 GB RAM	Results (Docker) Installed cleanly REPL in Connected Mode EVAL in connected Mode Deploy Demo Tutorial - first 2 examples EVAL, Deploy-Demo running together EVAL in standalone mode		Comments: If the path is wrong in eval mode, exception is not returned to user. Exception appears on server side. CORE-515 - Getting issue details... STATUS Casper issue, at some point the same block is passed around (observed at block 8) CORE-514 - Getting issue details... STATUS - Issue between connected and disconnected model
Mateusz	Ubuntu 16.04 (8 CPU 32GB RAM) Installation completes without issues Ctrl-C terminates the program without throwing an exception Program starts without exceptions (providing that the directories needed exist on the system) Started systemd service succesfully Started in standalone mode Tested REPL Tested EVAL Tested STDOUT Check Promethus server Check that all metrics in the release appear Server accepts requests and returns responses List of peers is accurate & matches what is presented on the server console. Response from peers reflects peers currently connected and changes as the peer list changes Deploy-demo call works (--deploy-demo) Comments: Still seeing NPE when I exit REPL with CTRL-C
Michael Birch	test locally with dev	assembly jar docker Result: Both work fine (modulo NPE on exit of repl via ^D)
Sebastian	test locally with dev	test docker with dev