Reduce default probe interval to 5 seconds (and fail-open timeout to 15).

Before now, the default probe interval (the idle time after which an echo
request is sent on an OpenFlow connection) was set to 15 seconds.  The
fail-open timeout is 3 times the probe interval, so this meant that it
took 45 seconds for a switch to fail open.

Users at Nicira have commented that this is too long.  They don't like the
idea that the network will be down for most of a minute before it begins to
recover.  So this commit changes the default probe interval to 5 seconds,
hence the fail-open timeout to 15 seconds.

diff --git a/debian/openvswitch-switch.template b/debian/openvswitch-switch.template
index c43def1..69f1de6 100644 (file)
--- a/debian/openvswitch-switch.template
+++ b/debian/openvswitch-switch.template
@@ -135,7 +135,7 @@ MGMT_VCONNS="punix:/var/run/secchan.mgmt"
 # INACTIVITY_PROBE: The maximum number of seconds of inactivity on the
 # controller connection before secchan sends an inactivity probe
 # message to the controller.  The valid range is 5 and up.  If unset,
-# secchan defaults to 15 seconds.
+# secchan defaults to 5 seconds.
 #INACTIVITY_PROBE=5
 
 # MAX_BACKOFF: The maximum time that secchan will wait between

diff --git a/secchan/ofproto.c b/secchan/ofproto.c
index 7445ef6..19baab9 100644 (file)
--- a/secchan/ofproto.c
+++ b/secchan/ofproto.c
@@ -333,7 +333,7 @@ ofproto_create(const char *datapath, const struct ofhooks *ofhooks, void *aux,
 
     /* Initialize OpenFlow connections. */
     list_init(&p->all_conns);
-    p->controller = ofconn_create(p, rconn_create(15, 8));
+    p->controller = ofconn_create(p, rconn_create(5, 8));
     p->controller->pktbuf = pktbuf_create();
     p->controller->miss_send_len = OFP_DEFAULT_MISS_SEND_LEN;
     p->listeners = NULL;

diff --git a/secchan/secchan.8.in b/secchan/secchan.8.in
index af6e9e8..2ad50ae 100644 (file)
--- a/secchan/secchan.8.in
+++ b/secchan/secchan.8.in
@@ -254,7 +254,7 @@ channel waits for a message to be received from the controller for
 controller.  After sending the inactivity probe, if no response is
 received for an additional \fIsecs\fR seconds, the secure channel
 assumes that the connection has been broken and attempts to reconnect.
-The default is 15 seconds, and the minimum value is 5 seconds.
+The default and the minimum value are both 5 seconds.
 
 When fail-open mode is configured, changing the inactivity probe
 interval also changes the interval before entering fail-open mode (see

diff --git a/vswitchd/bridge.c b/vswitchd/bridge.c
index e9c6390..b107320 100644 (file)
--- a/vswitchd/bridge.c
+++ b/vswitchd/bridge.c
@@ -1204,7 +1204,7 @@ bridge_reconfigure_controller(struct bridge *br)
         if (probe < 5) {
             probe = cfg_get_int(0, "mgmt.inactivity-probe");
             if (probe < 5) {
-                probe = 15;
+                probe = 5;
             }
         }
         ofproto_set_probe_interval(br->ofproto, probe);

diff --git a/vswitchd/ovs-vswitchd.conf.5.in b/vswitchd/ovs-vswitchd.conf.5.in
index 57af2aa..d613419 100644 (file)
--- a/vswitchd/ovs-vswitchd.conf.5.in
+++ b/vswitchd/ovs-vswitchd.conf.5.in
@@ -384,7 +384,7 @@ key.  If \fBovs\-vswitchd\fR does not communicate with the controller for the
 specified number of seconds, it will send a probe.  If a response is not
 received for an additional amount of that time, \fBovs\-vswitchd\fR assumes
 the connection has been broken and attempts to reconnect.  The default
-is 15 seconds, and the minimum value is 5 seconds.
+and minimum values are both 5 seconds.
 
 A management id may be specified with the \fBmgmt.id\fR key.  It takes
 an id in the form of exactly 12 hexadecimal digits.  If one is not