Common Error Code Table for CCE Container Network
Overview
This guide focuses on common troubleshooting scenarios in CCE container network. It provides explanations for event codes in network components and their corresponding solutions.
1. Pod error events
Pod error events primarily describe situations where pods fail to start normally. Network error event information of pods can be found on the pod object and all events related to pod can be queried via kubectl describe pod {podName}.
1.1 CNI errors
CNI errors occur when the container network component fails to create network resources for a pod during single-node pod creation. CCE categorizes pod creation failure reasons and provides respective solutions.
1.1.1 VPC-ENI mode
| Error code | Explanation | Trigger conditions | Solution |
|---|---|---|---|
| ENIIPCapacityExceed | The count of secondary IPs mounted on ENI has reached the quota limit | The count of secondary IPs on ENI created by the container network component exceeds the quota limit. Click to view ENI quota |
Add new nodes and assign pods to these new nodes. |
| ENICapacityExceed | ENI count has reached the quota limit | The ENI mounted on the node has reached its quota, and all secondary IPs bound to the ENI have also reached the quota limit. Click to view ENI quota |
Add new nodes and assign pods to these new nodes. |
| WaitCreateMoreENI | Wait for creation of new ENI | All secondary IPs of mounted ENIs on the node are exhausted; waiting for creation of new ENI | Wait for kubelet to automatically retry creating pod network resources |
| IPPoolExhausted | IPs in the cache pool are exhausted | The assignment of all IPs in the pre-requested IP address pool of the container network have been completed, waiting for pre-request of new IP by the container network component | Wait for kubelet to automatically retry creating pod network resources |
| SubnetNoMoreIP | The bound subnet IPs of the ENI have been exhausted | The bound subnet IPs for the ENI have been fully utilized, preventing the addition of new secondary IP addresses. | Solution 1. Create a larger subnet in the same availability zone as the node, and add the new subnet to the K8S Cluster CCE. Then delete the current node and purchase a new node in the same availability zone to join the K8S Cluster CCE. Solution 2. Use to allocate IPs via specified subnet. |
| failed to set bandwidth | The container network component cannot set bandwidth management for ENI | The container network component is unable to set bandwidth for a pod, possibly due to an outdated kernel version. | Contact customer service for resolution |
1.1.2 VPC-Route mode
| Error code | Explanation | Trigger conditions | Solution |
|---|---|---|---|
| NoMoreIP | Node has no available IP | All available IPs on the node have been used up, making it impossible to assign a new IP to the pod. | VPC-Route expansion container network segment |
2. Node error events
Network-related events in the CCE container network can be broadly categorized into two types: node errors and network resource errors.
2.1 Node error events
Node error events primarily describe situations where nodes fail to start normally. Network error event information of node can be found on the node object of K8s and all events related to node can be queried via kubectl describe node {nodeName}.
| Error code | Explanation | Trigger conditions | Solution |
|---|---|---|---|
| MetaAPIError01 | The container network component fails to obtain machine instances via MetaAPI | The CCE container network component was deployed on unsupported machine types | Contact customer service for resolution |
| MetaAPIError02 | The container network component fails to obtain machine region and other metadata information via MetaAPI | The CCE container network component was deployed on unsupported machine types | Contact customer service for resolution |
2.2 Network resource error events
Network resource error events primarily describe the cases where nodes fail to initialize or cannot complete resource preparation such as ENI. NetworkResourceSet is a network resource management object defined by CCE. You can query all network-related events associated with a node via kubectl describe nrs {nodeName}.
| Error code | Explanation | Trigger conditions | Solution |
|---|---|---|---|
| VPCQuotaLimitExceeded | Failed to create VPC route rule: Due to VPC resource quota limitations, no more VPC route rules can be created. Click to view VPC quota | VPC route rules have reached their upper limit | Submit a request in the quota center and contact customer service for resolution |
| CreateRouteRuleFailed | The container network component cannot create route via the open API | Experiencing sporadic issues when accessing the open API | CCE will automatically recover within 2 minutes. If it cannot recover for a long time, please contact the customer service |
| FailedCreateENI | The container network component cannot create ENI via the open API | Experiencing sporadic issues when accessing the open API | CCE will automatically recover within 2 minutes. If it cannot recover for a long time, please contact the customer service |
| NoAvailableSubnet | The availability zone where the node resides has no available subnet | The availability zone where the node resides has no available subnet, so the node cannot create a new ENI | Create a subnet in the same availability zone as the node, and add the new subnet to the K8S Cluster CCE |
| FailedBorrowSubnet | The subnet configured for the cluster does not exist | The subnet configured for the cluster is unavailable | Contact customer service for resolution |