[BUG] Cluster master election scheduler can stop after one renewal exception #6495

Current Behavior

Cluster master election/renewal can stop permanently after one exception escapes the scheduled task.

ShenyuClusterService.startSelectMasterTask() schedules master selection with scheduleAtFixedRate(...):

executorService.scheduleAtFixedRate(() -> doSelectMaster(host, port, contextPath), ...);

doSelectMaster() catches exceptions, closes the check services, and then throws a new ShenyuException:

} catch (Exception e) {
    LOG.error("select master error", e);
    upstreamCheckService.close();
    instanceCheckService.close();
    throw new ShenyuException(...);
}

For ScheduledThreadPoolExecutor, an exception escaping a fixed-rate task suppresses later executions of that periodic task. A transient lock/renewal failure can therefore stop future master selection attempts in that process.

Expected Behavior

Master election/renewal errors should be logged and the next scheduled election attempt should still run. Exceptions should not escape the periodic scheduler task.

Impact

After a transient cluster backend failure, the admin node may stop participating in master election until it is restarted.

Code Location

• shenyu-admin/src/main/java/org/apache/shenyu/admin/mode/cluster/service/ShenyuClusterService.java
  • startSelectMasterTask() schedules the periodic task.
  • doSelectMaster() rethrows from the catch block.