I'm seeing hangs for factory.clean() with Go IPFS nodes in the interop tests.
We used to have:
|
/** |
|
* Kill the `ipfs daemon` process. |
|
* |
|
* If the HTTP API is established, then send 'shutdown' command; otherwise, |
|
* process.kill(`SIGTERM`) is used. In either case, if the process |
|
* does not exit after 10.5 seconds then a `SIGKILL` is used. |
|
* |
|
* Note: timeout is ignored for proc nodes |
|
* |
|
* @param {Number} [timeout] - Use timeout to specify the grace period in ms before hard stopping the daemon. Otherwise, a grace period of 10500 ms will be used for disposable nodes and 10500 * 3 ms for non disposable nodes. |
|
* @returns {Promise} |
|
*/ |
|
killProcess (timeout) { |
Would be good to get a warning when force killing so we know that it is happening.