Implement transparent restart of components for multi-runner
Closed this issue · 1 comments
During our work with Crossbar at OpenDNA we have encountered scenarios where the TCPTransport associated with a Component dies, usually due to some issue with a WAMP operation the Component was engaged in failing in some as yet undetermined fashion (Currently suspected to be an issue with long-running or high-throughput operations running in parallel via asyncio.gather)
In the context of the multi-runner, this causes that specific Component to become disconnected rather than affecting every Component being managed by the multi-runner. To make recovery from this kind of issue simpler the multi-runner will be enhanced to allow transparent restarting of Components that become disconnected from the WAMP router
This issues was accidentally created in this repo when it should have been in https://github.com/opn-oss/autobahn-python-runners/issues/