From: Numan Siddique <nusid...@redhat.com>

The present code resets the database when it is in the state -
'RPL_S_SCHEMA_REQUESTED' and repopulates the database when it
receives the monitor reply when it is in the state -
'RPL_S_MONITOR_REQUESTED'. If however, it goes to active mode
before it processes the monitor reply, the whole data is lost.

This patch fixes the issue by resetting the database when it
receives the monitor reply (before processing it). so that
reset and repopulation of the db happens in the same state.

This approach still has a very small window for data loss if the
function process_notification() which processes the monitor
reply fails for some reason.

Reported-by: Han Zhou <zhou...@gmail.com>
Reported-at: 
https://mail.openvswitch.org/pipermail/ovs-discuss/2018-August/047161.html
Tested-by: aginwala <aginw...@ebay.com>
Signed-off-by: Numan Siddique <nusid...@redhat.com>
---
 ovsdb/replication.c | 20 ++++++--------------
 1 file changed, 6 insertions(+), 14 deletions(-)

diff --git a/ovsdb/replication.c b/ovsdb/replication.c
index 2b9ae2f83..44428a48e 100644
--- a/ovsdb/replication.c
+++ b/ovsdb/replication.c
@@ -299,19 +299,7 @@ replication_run(void)
                 /* After receiving schemas, reset the local databases that
                  * will be monitored and send out monitor requests for them. */
                 if (hmap_is_empty(&request_ids)) {
-                    struct shash_node *node, *next;
-
-                    SHASH_FOR_EACH_SAFE (node, next, replication_dbs) {
-                        db = node->data;
-                        error = reset_database(db);
-                        if (error) {
-                            const char *db_name = db->schema->name;
-                            shash_find_and_delete(replication_dbs, db_name);
-                            ovsdb_error_assert(error);
-                            VLOG_WARN("Failed to reset database, "
-                                      "%s not replicated.", db_name);
-                        }
-                    }
+                    struct shash_node *node;
 
                     if (shash_is_empty(replication_dbs)) {
                         VLOG_WARN("Nothing to replicate.");
@@ -335,7 +323,11 @@ replication_run(void)
             case RPL_S_MONITOR_REQUESTED: {
                 /* Reply to monitor requests. */
                 struct ovsdb_error *error;
-                error = process_notification(msg->result, db);
+                VLOG_INFO("Monitor request received. Resetting the database");
+                error = reset_database(db);
+                if (!error) {
+                    error = process_notification(msg->result, db);
+                }
                 if (error) {
                     ovsdb_error_assert(error);
                     state = RPL_S_ERR;
-- 
2.17.1

_______________________________________________
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to