Communication errors may temporarily occur in the network used as the admin network and log transfer network due to reasons such as high load on the server or insufficient system resources. Because of this, there is a risk of causing a split-brain situation by mistake even though the server has no issues.
Split brain is a phenomenon in which both servers temporarily operate as primary servers, causing data updates to be performed on both servers.
How to detect split brain using WebAdmin
If the conditions below are met, split brain may occur. Refer to "Split-brain detection method" and "How to recover from a split-brain" in "Appendix D Notes on Performing Automatic Degradation Immediately after a Heartbeat Abnormality" and take the actions described.
A standby instance is selected in the [Instances] tab, and
"Standalone" is displayed in [Instance type], and
A master instance is selected in the [Instances] tab, and
"Standalone" is displayed in [Instance type].
Note
The admin network is important because Mirroring Controllers use it to confirm the status of each server.
The log transfer network is also important to maintain the data freshness.
Therefore, use network configurations resistant to faults for these networks by using the network redundancy channel bonding feature provided by the operating system or network driver vendor.
