Message ID | 1486528822-38533-1-git-send-email-azhou@ovn.org |
---|---|
State | Accepted |
Headers | show |
On Tue, Feb 07, 2017 at 08:40:20PM -0800, Andy Zhou wrote: > Sometimes replication session can fail mostly due to replication > configurations. i.e. replicating from a database with a different > version of the schema. > > Currently, those errors are treated as fatal errors, and stops the > OVSDB server. A better way to handle those error may be to stop > only the replication session, and leave the OVSDB server up, so that > the replication can be restarted, may be with a different configuration, > at a later time. > > Signed-off-by: Andy Zhou <azhou@ovn.org> Acked-by: Ben Pfaff <blp@ovn.org>
diff --git a/ovsdb/ovsdb-server.c b/ovsdb/ovsdb-server.c index 1b0060b..9b669c9 100644 --- a/ovsdb/ovsdb-server.c +++ b/ovsdb/ovsdb-server.c @@ -199,8 +199,8 @@ main_loop(struct ovsdb_jsonrpc_server *jsonrpc, struct shash *all_dbs, if (*is_backup) { replication_run(); if (!replication_is_alive()) { - int retval = replication_get_last_error(); - ovs_fatal(retval, "replication connection failed"); + disconnect_active_server(); + *is_backup = false; } }
Sometimes replication session can fail mostly due to replication configurations. i.e. replicating from a database with a different version of the schema. Currently, those errors are treated as fatal errors, and stops the OVSDB server. A better way to handle those error may be to stop only the replication session, and leave the OVSDB server up, so that the replication can be restarted, may be with a different configuration, at a later time. Signed-off-by: Andy Zhou <azhou@ovn.org> --- ovsdb/ovsdb-server.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)