Based on our test results, in scenarios where the master node experiences 100% packet loss and process hang, it cannot switch quickly. Each probe takes about 1 minute to time out, which causes a 15 * 3 second switch to take 15 minutes to occur。
suggest setting the timeout time for each probe to 2 to 3 seconds, so that even in extreme scenarios, problems can be detected as early as possible and control can initiate the switch proactively。
Based on our test results, in scenarios where the master node experiences 100% packet loss and process hang, it cannot switch quickly. Each probe takes about 1 minute to time out, which causes a 15 * 3 second switch to take 15 minutes to occur。 suggest setting the timeout time for each probe to 2 to 3 seconds, so that even in extreme scenarios, problems can be detected as early as possible and control can initiate the switch proactively。