lambdaclass / reconnections Goto Github PK

The tests will work against services running in docker.

• There needs to be a way (perhaps calling os:cmd) to interface with the running docker to tell it to kill a service.
• There needs to be a way to ask docker or the os whether a service is running
• There needs to be some process randomly killing (and resurrecting) services during the test run
• The test should constantly try to use the service.
• when the service call fails check that both the OS and the driver report the service as being disconnected
• after a disconnection, test that the driver attempts reconnection according to the configured backoff, that reconnection succeeds and that service calls work again.

docker run --rm -it -v (pwd):/reconnections -v ~/.cache:/root/.cache -w /reconnections erlang rebar3 as test shell

don't use applications:get_env/2 in gen_server

%% priv

The servers always do the same, driver should only encapsulate whatever changes from driver to driver.

Consider making driver a behavior.

• Enter before same functions with different parameters
• No ASCII separators
• Rename server/ to drivers/
• Use Maps instead of touples in supervisor

• epgsql
• eredis
• cqerl
• riak erlang client
• emysql
• rabbitMq
• brod
• MongoDB Erlang
• neo4j
• sql lite
• Amazon DynamoDB
• CouchDB