Very common to allow the Aurora Serverless autoPause after some period of time. When this happens and Rails tries to connect, you will see an error like this pretty quickly.
Aws::RDSDataService::Errors::BadRequestException (Communications link failure)
The last packet sent successfully to the server was 0 milliseconds ago. The driver has not received any packets from the server.
At one time I had code in place to handle this but removed it due to the issues from #8 for flaky tests because I assumed it was due to the Retriable (https://github.com/kamui/retriable) gem. It was not, so it is safe to add this back. Here is quick implementation I had that we can add back. I liked this method since it meant we only used retrys for the initial connect and not all the time for each execute statement.
def execute_statement_retry(sql)
if @connected
execute_statement(sql)
else
auto_paused_retry { execute_statement(sql) }
end
end
def execute_statement(sql)
raw_client.execute_statement({
# ...
}).tap do |r|
@connected = true
# ...
end
end
private
def auto_paused_retry
ekls = Aws::RDSDataService::Errors::BadRequestException
emsg = /last packet sent successfully to the server was/
rtry = Proc.new { Rails.logger.info 'Aurora auto paused, retrying...' }
optn = { on: { ekls => emsg }, on_retry: rtry, tries: 10 }
Retriable.retriable(optn) { yield }
end
Very common to allow the Aurora Serverless
autoPauseafter some period of time. When this happens and Rails tries to connect, you will see an error like this pretty quickly.At one time I had code in place to handle this but removed it due to the issues from #8 for flaky tests because I assumed it was due to the Retriable (https://github.com/kamui/retriable) gem. It was not, so it is safe to add this back. Here is quick implementation I had that we can add back. I liked this method since it meant we only used retrys for the initial connect and not all the time for each execute statement.