It can happen that both nodes break down simultaneously (for example because of a power failure), or the slave node breaks down before the original master node recovers. The following describes how to properly recover syslog-ng Store Box (SSB).
NOTE: When both nodes of a cluster boot up in parallel, the node with the 1.2.4.1 HA IP address will become the master node.
To properly recover SSB
-
Power off both nodes by pressing and releasing the power button.
Caution: Hazard of data loss If SSB does not shut down, press and hold the power button for approximately 4 seconds. However, consider that this method terminates connections passing SSB and might result in data loss.
-
Power on the node that was the master before SSB broke down. Check the system logs to find out which node was the master before the incident: when a node boots as master, or when a takeover occurs, SSB sends a log message identifying the master node.
TIP: Configure remote logging to send the log messages of SSB to a remote server where the messages are available even if the logs stored on SSB become inaccessible. For details on configuring remote logging, see SNMP and email alerts.
-
Wait until this node finishes the boot process.
-
Power on the other node.
