mds: replace closed sessions on connect

If a connection comes and there is a closed session attached, remove it.
This is probably a failure of an old session to get cleaned up properly,
and in certain cases it may even be from a different client (if the addr
nonce is reused).  In that case this prevents further damage, although
a complete solution would also clean up the closed connection state if
there is a fault.  See #3630.

This fixes a hang that is reproduced by running the libcephfs
Caps.ReadZero test in a loop; eventually the client addr is reused and
we are linked to an ancient Session with a different client id.

Backport: bobtail
Signed-off-by: Sage Weil <sage@inktank.com>

diff --git a/src/mds/MDS.cc b/src/mds/MDS.cc
index cbcda25a8dd7812fffece703c87208947ec384a1..211fd5c6e8aa1b68b6f88c9f4e61d3315aaef633 100644 (file)
--- a/src/mds/MDS.cc
+++ b/src/mds/MDS.cc
@@ -2056,6 +2056,12 @@ bool MDS::ms_verify_authorizer(Connection *con, int peer_type,
     // wire up a Session* to this connection, and add it to the session map
     entity_name_t n(con->get_peer_type(), global_id);
     Session *s = sessionmap.get_session(n);
+    if (s && s->is_closed()) {
+      dout(10) << " have session " << s << " but it is CLOSED, removing" << dendl;
+      con->set_priv(NULL);
+      s->put();
+      s = NULL;
+    }
     if (!s) {
       s = new Session;
       s->inst.addr = con->get_peer_addr();