FAQ about the service console - Realtime Compute for Apache Flink

This topic provides answers to some frequently asked questions about the service console.

What are the meanings of different colors for the legends in the Status Durations column on the Subtask Metrics tab?
How do I calculate the number of consumed CUs in a workspace?
If the memory size of the JVM metaspace in a session cluster in which a deployment is deployed is always high, does the metaspace occupy memory resources after I cancel the deployment?
What do I do if the error message "One bucket not allowed exceed ten items of CORS Rules" appears?
What do I do if the error message "Has not enough ip address: abnormal event detected from kubernetes" appears?

What are the meanings of different colors for the legends in the Status Durations column on the Subtask Metrics tab?

The value in the Status Durations column indicates the duration of each phase of the vertex subtask. This section describes the meaning of the color of each node box.

: CREATED
: SCHEDULED
: DEPLOYING
: INITIALIZING
: RUNNING

How do I calculate the number of consumed CUs in a workspace?

On the O&M > Deployments page, view the number of CPU cores and the memory size of all deployments in each namespace.
On the O&M > Session Clusters page, view the number of CPU cores and the memory size of all session clusters in each namespace.
Calculate the sum of CPU cores and the total memory size in the workspace to obtain the total number of CPU cores and the total memory size in each namespace.
Calculate the number of consumed compute units (CUs) in a workspace by using the following formula: MAX (Total number of CPU cores, Total memory size/4).
The value of the Consumed /Purchased CUs parameter in the Realtime Compute for Apache Flink console indicates the number of consumed CUs that you obtained. The following figure shows how to view the value of the Consumed /Purchased CUs parameter.

If the memory size of the JVM metaspace in a session cluster in which a deployment is deployed is always high, does the metaspace occupy memory resources after I cancel the deployment?

In most cases, the metaspace is released after a deployment is canceled. However, the release may be delayed for several minutes. Some threads of tasks that are periodically run in the JobManager use objects of a deployment. If an object of the deployment is not recycled, the classloader of the deployment cannot be released and the memory that is related to all classes loaded in the classloader cannot be released.

What do I do if the error message "One bucket not allowed exceed ten items of CORS Rules" appears?

Problem description
Cause
A maximum of 10 cross-origin resource sharing (CORS) rules for an Object Storage Service (OSS) bucket can be created. When you upload a file in the development console of Realtime Compute for Apache Flink, OSS automatically creates a CORS rule. As a result, the number of CORS rules exceeds the upper limit.
Solution
Delete or merge some CORS rules in OSS. For more information, see Use the OSS console.

What do I do if the error message "Has not enough ip address: abnormal event detected from kubernetes" appears?

Problem description

Has not enough ip address:abnormal event detected from kubernetes (type:[Warning], reason:[CniError_CodeUnKnownErr], message:[CniAllocateError: allocateIP failed: ipamCreate failed: failed to create ENI: all vSwitches ([*****]) cannot be used: CreateNetworkInterface: RequestId: 67959AE5-EA20-5CB4-8560-5BD6752472FD, ErrorCode: InvalidVSwitchId.IpNotEnough, Message: The specified VSwitch "*****" has not enough IpAddress., elapsedTime: 245.03232ms])

Cause
No IP address is available in the CIDR block of the vSwitch.
Solution
Create a vSwitch that has an available IP address in the Realtime Compute for Apache Flink console. For more information, see Modify a vSwitch.