4 ms·
One thing missing from this article is TCP heartbeats, which prevents the first two conditions from going on indefinitely. However since the minimum allowed he
by syncsynchalt 10y ago
One thing missing from this article is TCP heartbeats, which prevents the first two conditions from going on indefinitely.
However since the minimum allowed heartbeat is (usually) two hours host reboot/shutdown is usually detected via a heartbeat at the application layer instead, which sends some no-op traffic every n seconds or minutes.
- betenoire 10y agolike you say, the heartbeat removes some instances, but not others. The article also says, "It's the responsibility of the application to handle the cases where these differ (often using a keep-alive mechanism)."
- ghusbands 10y agoThe article mentions both. "It's possible to manage this problem using TCP or application-level keep-alive." (and did before you commented)