cluster: STABLE3 - dlm_controld: add more fs_notified debugging

David Teigland teigland at fedoraproject.org
Tue Mar 23 20:59:20 UTC 2010


Gitweb:        http://git.fedorahosted.org/git/cluster.git?p=cluster.git;a=commitdiff;h=7b8f7da923c5f18e6c5e688ee6a2db981e51f22d
Commit:        7b8f7da923c5f18e6c5e688ee6a2db981e51f22d
Parent:        7fde51f8331e11e1a8cfcaa00bc90c9cc5d420f5
Author:        David Teigland <teigland at redhat.com>
AuthorDate:    Tue Mar 23 15:57:48 2010 -0500
Committer:     David Teigland <teigland at redhat.com>
CommitterDate: Tue Mar 23 15:59:14 2010 -0500

dlm_controld: add more fs_notified debugging

To help figure out what's gone wrong if/when an fs_controld
goes into an fs_notified loop, which we see from time to time.

bz 576335

Signed-off-by: David Teigland <teigland at redhat.com>
---
 group/dlm_controld/cpg.c |    6 +++++-
 1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/group/dlm_controld/cpg.c b/group/dlm_controld/cpg.c
index e01ecc2..20d59d5 100644
--- a/group/dlm_controld/cpg.c
+++ b/group/dlm_controld/cpg.c
@@ -506,8 +506,10 @@ static void node_history_fail(struct lockspace *ls, int nodeid,
 	if (cfgd_enable_quorum && !cfgd_enable_fencing)
 		node->check_quorum = 1;
 
-	if (ls->fs_registered)
+	if (ls->fs_registered) {
+		log_group(ls, "check_fs nodeid %d set", nodeid);
 		node->check_fs = 1;
+	}
 
 	node->removed_seq = cg->seq;	/* for queries */
 	node->failed_reason = reason;	/* for queries */
@@ -635,6 +637,7 @@ static int check_fs_done(struct lockspace *ls)
 			continue;
 
 		if (node->fs_notified) {
+			log_group(ls, "check_fs nodeid %d clear", node->nodeid);
 			node->check_fs = 0;
 		} else {
 			log_group(ls, "check_fs nodeid %d needs fs notify",
@@ -2344,6 +2347,7 @@ int set_fs_notified(struct lockspace *ls, int nodeid)
 		return -EAGAIN;
 	}
 
+	log_group(ls, "set_fs_notified nodeid %d", nodeid);
 	node->fs_notified = 1;
 	return 0;
 }


More information about the cluster-commits mailing list