jmsperu commented on PR #13090: URL: https://github.com/apache/cloudstack/pull/13090#issuecomment-6372731851
@sudo87 thanks for the review. On the two points in the review body: **1. Stale description.** Updated the description. It now covers what each action does to the VMs (`restart-agent` and `log-only` stop them first, `graceful-reboot` leaves it to the shutdown, `hard-reboot` does not touch them), all five properties with defaults (`action`, `keep.vms.running`, `include.system.vms`, `stop.vms.timeout` and the new `restart.agent.cooldown`), the deadman and its timing, the fallback chain, and the two lab runs. **2. restart-agent loop.** Added a guard in 7c3ff67cd3. Each `restart-agent` fence writes its time to `/run/cloudstack-agent/kvmha-fence/restart-agent.last` just before restarting the agent; `/run` survives the agent restart but not a reboot. If `restart-agent` fires again within `kvm.heartbeat.fence.restart.agent.cooldown` seconds (new property, default 600, `0` turns it off), the fence logs why and escalates to hard-reboot. If the marker cannot be written it also hard-reboots, since a loop could not be detected. Documented in agent.properties and AgentProperties. Tests: `restartAgentRecordsWhenItRestartedTheAgent`, `restartAgentWithinTheCooldownEscalatesToHardReboot`, `restartAgentAfterTheCooldownRestartsTheAgentAgain`, `restartAgentCooldownZeroTurnsTheGuardOff`, `invalidRestartAgentCooldownFallsBackToTheDefault`. The inline threads have their own replies. Head is now 7c3ff67cd3. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
