clear cluster_master on election so that slaves will accept a new master

[l2tpns.git] / Changes

diff --git a/Changes b/Changes
index 12cb3f8..0df180a 100644 (file)
--- a/Changes
+++ b/Changes
@@ -1,4 +1,17 @@
-* Sat May 21 2005 Brendan O'Dea <bod@optusnet.com.au> 2.0.21
+* Sun Jul 3 2005 Brendan O'Dea <bod@optus.net> 2.0.23
+- Backout regular_cleanup changes.
+- Clear cluster_master on election so that slaves will accept a new master.
+
+* Tue May 31 2005 Brendan O'Dea <bod@optusnet.com.au> 2.0.22
+- Show session open time in "show session"/"show user" detailed output.
+- Have slaves with BGP configured drop BGP on receipt of a shutdown
+  signal, but hang about for an additional 5s to process any remaining
+  traffic.
+- Run regular_cleanups after processing the results of the select,
+  looking at a sufficient slice of each table to ensure that all
+  entries are examined at least once per second.
+
+* Mon May 23 2005 Brendan O'Dea <bod@optusnet.com.au> 2.0.21
 - Cluster changes from Michael, intended to prevent a stray master
   from trashing a cluster:
   + Ignore heartbeats from peers claiming to be the master before the
@@ -6,12 +19,16 @@
   + A master receiving a stray heartbeat sends a unicast HB back, which
     should cause the rogue to die due to the tie-breaker code.
   + Keep probing the master for late heartbeats.
-  + Drop BGP as soon as we become master with peers (TODO: pre-emptively
-    drop routes if the master is late and we are the best candidate).
-  + A slave (presumably a recently restarted master) receiving a
-    LASTSEEN message sends a zero basetime ping (as sent by the master
-    on clean shutdown).
-  
+  + Drop BGP as soon as we become master with the minumum required peers.
+  + Any PING seen from a master forces an election (rather than just
+    where basetime is zero).
+  + A slave which receives a LASTSEEN message (presumably a restarted
+    master) sends back new message type, C_MASTER which indicates the
+    address of the current master.
+- New config option: cluster_master_min_adv which determines the minimum
+  number of up to date slaves required before the master will drop
+  routes.
+
 * Mon May 16 2005 Brendan O'Dea <bod@optusnet.com.au> 2.0.20
 - Add handling of "throttle=N" RADIUS attributes.
 - Fix RADIUS indexing (should have 16K entries with 64 sockets).