You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the coordinator uses a mix ot hackery and tricks to compute its IP address. This involves a call to gethostname() followed by a call to getaddrinfo(). This is still unreliable and can give us 127.0.0.1 as the IP address, when we actually need the non-loopback IP.
The text was updated successfully, but these errors were encountered:
@karya0 127.0.0.1 is given in the scenario where dmtcp_launch has no --host parameter. At this point, SharedData is not initialized. If ENV_VAR_NAME_HOST is not set, hostname is set to DEFAULT_HOST, which is 127.0.0.1. _coordinatorSocket is then created using this address, which is wrong. This is harmless when running a local computation. But for a distributed one, remote processes won't connect to the correct coordinator.
Currently, the coordinator uses a mix ot hackery and tricks to compute its IP address. This involves a call to gethostname() followed by a call to getaddrinfo(). This is still unreliable and can give us 127.0.0.1 as the IP address, when we actually need the non-loopback IP.
The text was updated successfully, but these errors were encountered: