Fix 'test_cluster_sync-t' assumption on 'mysql_servers' status

pull/4161/head
Javier Jaramago Fernández 3 years ago
parent 3f79710878
commit 44098b0d79

@ -2255,6 +2255,20 @@ int main(int, char**) {
sleep(2);
// NOTE: Recovering the DISK configuration shouldn't be required. But due to current limitations
// regarding monitor a server could be permanently moved from it's original hostgroup in user
// configuration (mysql_servers table). A scenario like this could for example be:
// - A server is moved by Monitoring actions during any of the previous sync tests for hostgroups
// tables, for example, placed in the OFFLINE_HOSTGROUP.
// - A later reconfiguration via 'read_only_action' rewrites Admin 'mysql_servers' table, making this
// server permanent in user config.
// - The following checks expects to find this server in a particular hostgroup. But config is
// permanently altered, and fail.
// The possibility of this scenario make the backup mechanism of the previous sections insufficient. So
// right now the safest option is recover DISK configuration.
MYSQL_QUERY(proxy_admin, "LOAD MYSQL SERVERS FROM DISK");
MYSQL_QUERY(proxy_admin, "LOAD MYSQL SERVERS TO RUNTIME");
// Check sync disable via 'admin-cluster_*_sync' variables
{
int checksum_sync_res = check_modules_checksums_sync(proxy_admin, r_proxy_admin, cl);

Loading…
Cancel
Save