Troubleshoot AI Pipelines¶
This section helps you identify and solve some issues when using AI Pipelines.
“Duplicate Key Entry” error when running the MySQL Writer operator.¶
A Duplicate Key Entry
error occurs when you run an MySQL Writer operator.
Cause¶
The value of record time is duplicated due to the automatic update of current time.
Solution¶
Add record_time = record_time
as an input parameter to ensure the value is constant.
Fail to run the “pip install” command in a Python operator¶
You fail to run a pipeline instance, where the log of the Python operator records an error related to pip install
command.
Cause¶
This issue occurs due to insufficient environmental requirements.
Solutions¶
Solutions 1: check your network connection, and try again.
Solutions 2: use custom operators to fix the issue. For more information on custom operators, see Custom Operators.
“Request entity too large:limit is 3145728” error when running a Python operator¶
A request entity too large:limit is 3145728
error occurs when you try to run a Python operator.
Cause¶
You refer to the output parameters of other operators as the input parameter of this Python operator in list or string type, but the size of output data is over 3M.
Solution¶
Solution 1: refer to output parameters in directory or file type.
Solution 2: use external storage for output reference, such as S3 or MySQL.
“Error:root:input format parse failed” error when running a Model operator¶
An Error:root:input format parse failed
error occurs when you try to run a Model operator.
Cause¶
The range
parameter is undeclared in the input_format declaration of the Model operator.
Solution¶
Configure range
in the input_format declaration of the Model operator. An example of range
looks like this:
"range": [
0,
440
],
“Error:root:miss parameterrange” error when running a Single Model Deployment operators¶
An Error:root:miss parameterrange
error occurs when you try to run a Single Model Deployment operator.
Cause¶
The output_format and/or input_format parameters are not in json format.
Solution¶
Make sure the values of output_format and input_format parameters are in json format. An example of input_format in json looks like this:
[
{
"annotations": "",
"defaultValue": 300,
"dtype": "int",
"ftype": "continuous",
"name": "i.set",
"range": [
0,
440
],
"repeat": 0
},
]
“Modeldeployments.eap.envision.com "xxx-deployment" not found” error when running a Model Instance operator¶
A modeldeployments.eap.envision.com \"xxx-deployment\" not found
error occurs when you try to run a Model Instance operator.
Cause¶
This issue occurs because the resource pool of the pipeline is incorrect.
Solution¶
Select the resource pool starting with mmc-
. An example of the correct resource pool looks like this:
mmc-o16062946385871018
“Failed with exit code 137” error when running an operator¶
A failed with exit code 137
error occurs when you try to run an operator.
Cause¶
The issue occurs due to full physical node memory.
Solution¶
In AI pipeline canvas, go to Workflow setting > Advanced Config.
Increase the value of Memory request and Memory limit.
“Create unit [{operator1}, {operator2}, {operator3}] failed. Create container {container1} failed. Failed to find the operator item param info “ error when running a pipeline canvas¶
A create unit [{operator1}, {operator2}, {operator3}] failed. Create container {container1} failed. Failed to find the operator item param info
error occurs when you try to run a pipeline canvas.
In which:
{Operator1}
,{operator2}
, and{operator3}
refers to the operators in the pipeline canvas.{Container1}
refers to the canvas you try to run.
Cause¶
You refer to the parameters that do not exist as the input parameters for the operators mentioned in the error message.
Solution¶
Double-check the input parameter configuration of the operators mentioned in the error message to ensure you refer to the correct parameters.