[DRBD-cvs] svn commit by phil - r2207 - trunk - Documented the idea how to synchronise the start of res

[drbd-cvs at lists.linbit.com]

Author: phil
Date: 2006-05-22 22:02:45 +0200 (Mon, 22 May 2006)
New Revision: 2207

Modified:
   trunk/ROADMAP
Log:
Documented the idea how to synchronise the start of resync.


Modified: trunk/ROADMAP
===================================================================
--- trunk/ROADMAP	2006-05-22 18:58:45 UTC (rev 2206)
+++ trunk/ROADMAP	2006-05-22 20:02:45 UTC (rev 2207)
@@ -727,11 +727,14 @@
    Connected -> SyncSource in the cluster.
 
       role <- primary
-      conn <- SyncSource
-      conn <- SyncTarget
+      disk <- inconsistent (conn <- SyncTarget) **
+      pdsk <- inconsistent (conn <- SyncSource)
       disk <- Diskless (as long as it happens as administrative command)
       pdsk <- Outdated (= a 'disconnect' issued on a primary node)
 
+   ** It is more feasable to make the change on disk/pdsk cluster
+      wide atomic than working with the connection state.
+
    * When a state change might sleep ( reuqest_state() ) and it is 
      to be cluster wide atomic ( pre_state_checks() determines this!).
 	1. Aquire the cluster state change lock (bit & waitqueue) ?