Allow ports to be reused in gloo (#97677)

Summary:
X-link: pytorch/pytorch#97677

Pull Request resolved: #353

ProcessGroupGloo and gloo seem to be opening and closing sockets without allowing the port to be reused. We see this issue pop up in larger training jobs "Address already in use" and we assume it to be because all the ephemeral ports are exhausted.

This diff allows ports to be reused, we see a reduced number of ports being in `TIME_WAIT` state.

context: https://fb.workplace.com/groups/319878845696681/permalink/5988899781205532/

another issue: https://fb.workplace.com/groups/319878845696681/permalink/958768178474408/

Differential Revision: D44029927

fbshipit-source-id: b531d67456d4656ce23e9db7c4cb892d8fc90475

gloo/transport/tcp/device.cc

@@ -101,6 +101,15 @@ static void lookupAddrForHostname(struct attr& attr) {
   struct addrinfo* rp;
   for (rp = result; rp != nullptr; rp = rp->ai_next) {
     auto fd = socket(rp->ai_family, rp->ai_socktype, rp->ai_protocol);
+
+    // Set SO_REUSEADDR to signal that reuse of the listening port is OK.
+    int on = 1;
+    rv = setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, reinterpret_cast<const char*>(&on), sizeof(on));
+    if (rv == -1) {
+      close(fd);
+      GLOO_ENFORCE_NE(rv, -1);
+    }
+
     if (fd == -1) {
       continue;
     }