Hi team!
Here's a case that After an OceanStor V3 Storage System with the HyperMetro Feature Is Upgraded, the Historical Event List Contains an Alarm Indicating that a Controller Cannot Be Monitored.
Problem Description
After an OceanStor V3 storage system with the HyperMetro feature is upgraded (from V300R003C00 to V300R003C00SPC100), the historical event list contains an alarm indicating that a controller cannot be monitored.
In DeviceManager, the current alarm is cleared.
Symptom
The storage system upgraded to V300R003C00SPC100 reports the following alarm:
The controller module (Controller Enclosure CTE0, id B) cannot be monitored.
The alarm ID is 0xF00CF0014.
Possible Cause
When V300R003C00 interworks with V300R003C00SPC100, the inter-array replication link environment is as follows:
One controller is connected through two links based on parallel networking and a HyperMetro domain is configured.
If all inter-array replication links are disconnected and then simultaneously recovered,
the V300R003C00's replication link heartbeats simultaneously resume, and recovery events about two links are generated. As a result, V300R003C00 initiates two identical version negotiations to V300R003C00SPC100 at the same time.
Packet processing is abnormal, leading to controller Oops.
Identification Method
1. Two storage systems configured as a HyperMetro pair are upgraded from V300R003C00 to V300R003C00SPC100.
2. The inter-array replication link environment is as follows: One controller is connected through two links based on parallel networking and a HyperMetro domain is configured.
3. The primary controller of the storage system that is upgraded successfully has the Oops problem.
Use Toolkit to collect logs and find out the power-on logs generated near the alarm time in the following directory:
The kbox_die_1.txt file contains the following stack information:
In the following directory, find out the log_debug.txt file that is generated near the alarm time:
Search for the keyword versServReadReqTlvHandle.
The following error-level EPL log information exists, indicating that requesting a buffer for version negotiation fails.
Workaround
Before performing an upgrade, check whether links are connected between storage systems based on the parallel networking mode.
If yes, change the parallel networking mode to the HyperMetro standard networking mode under the condition that services are not affected.
Parallel networking
Standard networking
