Node error codes describe failures that relate to a specific node canister.
Because node errors are specific to a node, for example, memory has failed, the errors are only reported on that node. However, some of the conditions that the node detects relate to the shared components of the enclosure. In these cases both node canisters in the enclosure report the error.
There are two types of node errors: critical node errors and noncritical node errors.
A critical error means that the node is not able to participate in a clustered system until the issue that is preventing it from joining a clustered system is resolved. This error occurs because part of the hardware has failed or the system detects that the scode is corrupt. If it is possible to communicate with the canister with a node error, an alert that describes the error is logged in the event log. If the system cannot communicate with the node canister, a Node missing alert is reported. If a node has a critical node error, it is in service state, and the fault LED on the node is on. The exception is when the node cannot connect to enough resources to form a clustered system. It shows a critical node error but is in the starting state. The range of errors that are reserved for critical errors are 500–699.
A noncritical error code is logged when there is a hardware or code failure that is related to just one specific node. These errors do not stop the node from entering active state and joining a clustered system. If the node is part of a clustered system, there is also an alert that describes the error condition. The node error is shown to make it clear which of the node canisters the alert refers to. The range of errors that are reserved for noncritical errors are 700–899.