[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

[Cluster-devel] Cluster Project branch, STABLE, updated. rgmanager_1_9_66-12-g5ee43ba



This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "Cluster Project".

http://sources.redhat.com/git/gitweb.cgi?p=cluster.git;a=commitdiff;h=5ee43bab4b9c3048e58eb5f1de61d57b1a28e045

The branch, STABLE has been updated
       via  5ee43bab4b9c3048e58eb5f1de61d57b1a28e045 (commit)
       via  5459b747db65955eb0ddced1524e909e0d19f52f (commit)
      from  67f3b6dd2c711cb1077b88daea7bb6dd27fcf0ec (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit 5ee43bab4b9c3048e58eb5f1de61d57b1a28e045
Author: Lon Hohberger <lhh redhat com>
Date:   Fri Jan 25 18:13:24 2008 +0000

    Fix qdiskd master abdication logic (#430264)

commit 5459b747db65955eb0ddced1524e909e0d19f52f
Author: Lon Hohberger <lhh redhat com>
Date:   Tue Mar 11 10:43:37 2008 -0400

    [CMAN] Fix "Node X is undead" loop bug
    
    This was caused by an improper assignment to ps_incarnation
    after a node decides to evict another node.  The fix
    is to simply make the internal (memory) assignments before
    calling qd_write_status()

-----------------------------------------------------------------------

Summary of changes:
 cman/qdisk/main.c |   37 +++++++++++++++++++------------------
 1 files changed, 19 insertions(+), 18 deletions(-)

diff --git a/cman/qdisk/main.c b/cman/qdisk/main.c
index 1956c46..cb20590 100644
--- a/cman/qdisk/main.c
+++ b/cman/qdisk/main.c
@@ -252,23 +252,6 @@ check_transitions(qd_ctx *ctx, node_info_t *ni, int max, memb_mask_t mask)
 		     state_run(ni[x].ni_status.ps_state)) {
 
 			/*
-			   Write eviction notice if we're the master.
-			 */
-			if (ctx->qc_status == S_MASTER) {
-				clulog(LOG_NOTICE,
-				       "Writing eviction notice for node %d\n",
-				       ni[x].ni_status.ps_nodeid);
-				qd_write_status(ctx, ni[x].ni_status.ps_nodeid,
-						S_EVICT, NULL, NULL, NULL);
-				if (ctx->qc_flags & RF_ALLOW_KILL) {
-					clulog(LOG_DEBUG, "Telling CMAN to "
-						"kill the node\n");
-					cman_kill_node(ctx->qc_ch,
-						ni[x].ni_status.ps_nodeid);
-				}
-			}
-
-			/*
 			   Mark our internal views as dead if nodes miss too
 			   many heartbeats...  This will cause a master
 			   transition if no live master exists.
@@ -285,6 +268,23 @@ check_transitions(qd_ctx *ctx, node_info_t *ni, int max, memb_mask_t mask)
 			ni[x].ni_evil_incarnation = 
 				ni[x].ni_status.ps_incarnation;
 			
+			/*
+			   Write eviction notice if we're the master.
+			 */
+			if (ctx->qc_status == S_MASTER) {
+				clulog(LOG_NOTICE,
+				       "Writing eviction notice for node %d\n",
+				       ni[x].ni_status.ps_nodeid);
+				qd_write_status(ctx, ni[x].ni_status.ps_nodeid,
+						S_EVICT, NULL, NULL, NULL);
+				if (ctx->qc_flags & RF_ALLOW_KILL) {
+					clulog(LOG_DEBUG, "Telling CMAN to "
+						"kill the node\n");
+					cman_kill_node(ctx->qc_ch,
+						ni[x].ni_status.ps_nodeid);
+				}
+			}
+
 			/* Clear our master mask for the node after eviction */
 			if (mask)
 				clear_bit(mask, (ni[x].ni_status.ps_nodeid-1),
@@ -932,7 +932,8 @@ quorum_loop(qd_ctx *ctx, node_info_t *ni, int max)
 		ctx->qc_master = master_exists(ctx, ni, max, &low_id, &count);
 
 		/* Resolve master conflict, if one exists */
-		if (count > 1 && ctx->qc_status == S_MASTER) {
+		if (count >= 1 && ctx->qc_status == S_MASTER &&
+		    ctx->qc_master != ctx->qc_my_id) {
 			clulog(LOG_WARNING, "Master conflict: abdicating\n");
 
 			/* Handle just like a recent upgrade */


hooks/post-receive
--
Cluster Project


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]