Description
I’m following the RAC container setup documentation using the setup_rac_host.sh script to spin up RAC containers via Podman.
Environment details:
- Platform: Google Cloud Platform (GCP) VM
- OS: Oracle Linux 8
Reference: OracleDatabase/RAC/OracleRealApplicationClusters
The script successfully provisions the RAC nodes (racnodep1, racnodep2), but during the cluster startup phase, I encounter repeated gipcretConnectionLost and no network HB messages, indicating the interconnect heartbeat is failing.
Observed Logs
Below is a snippet of the recurring error messages from racnodep2:
2025-10-24 17:46:42.320 : ONMD: [INFO] clssnmvDHBValidateNCopy: node 1, racnodep1, has a disk HB, but no network HB
2025-10-24 17:46:42.391 :GIPCHAUP: [INFO] gipchaUpperProcessDisconnect: processing DISCONNECT for hendp 0x7f140c03e570
2025-10-24 17:46:42.391 :GIPCGMOD: [INFO] gipcmodGipcCallbackDisconnect: Disconnect forced for endp 0x7f140c042c10
2025-10-24 17:46:43.391 : ONMD: [INFO] clssnmDiscHelper: racnodep1 connection failed, endp (0x9d7)
Expected Behavior
RAC containers should establish stable interconnect communication (gipcha and clssnm layers) between nodes when configured through setup_rac_host.sh.
Request
Could the maintainers please:
- Confirm if RAC-on-Podman is supported on Oracle Linux 8 under GCP?
- Suggest any required network adjustments (bridge, MACVLAN, multicast) for the interconnect?