finally subtask no ignoreError
Showcases that you can can do finally cleanup and resuce the panic in the sub task level
Demo
source
Main task yaml file
tasks:
- name: task
task:
- func: call
desc: task
do:
- sub_task_layer1
- name: sub_task_layer1
desc: sub_task_layer1
task:
- func: call
loop:
- item1
- item2
do:
- sub_task_layer2
- name: sub_task_layer2
task:
- func: shell
name: step1
desc: step 1
do:
- 'echo """opening file: {{.loopitem}}"""'
- echo "hello"|grep "world"
- func: cmd
name: step2
desc: |
in this case, since there is no ignoreError, the exception was captured by task level finaly code block
opened file is safely closed
to make the flow to continue to reach step2, use ignoreError
do:
- name: print
cmd: step 2
desc: |
without rescue, the execution will return a non-zero return code in shell and also report the error
with rescue, the program will return 0
rescue: true
finally:
- func: shell
name: close_file
desc: |
ensure the opened file is closed
do:
- echo "close the file ....."
Main log file
loading [Config]: ./tests/functests/upconfig.yml
Main config:
Version -> 1.0.0
RefDir -> ./tests/functests
WorkDir -> cwd
AbsWorkDir -> /up_project/up
TaskFile -> f0185
Verbose -> vvv
ModuleName -> self
ShellType -> /bin/sh
MaxCallLayers -> 8
Timeout -> 3600000
MaxModuelCallLayers -> 256
EntryTask -> task
ModRepoUsernameRef ->
ModRepoPasswordRef ->
work dir: /up_project/up
-exec task: task
loading [Task]: ./tests/functests/f0185
module: [self], instance id: [dev], exec profile: []
profile - envVars:
(*core.Cache)({
})
Task1: [task ==> task: ]
-Step1: [: task ]
self: final context exec vars:
(*core.Cache)({
"up_runtime_task_layer_number": 0
})
=Task2: [task ==> sub_task_layer1: sub_task_layer1 ]
--Step1:
self: final context exec vars:
(*core.Cache)({
"up_runtime_task_layer_number": 1
})
==Task3: [task/sub_task_layer1 ==> sub_task_layer2: without rescue, the execution will return a non-zero return code in shell and also report the error
with rescue, the program will return 0
]
---Step1: [step1: step 1 ]
self: final context exec vars:
(*core.Cache)({
"up_runtime_task_layer_number": 2,
"loopitem": "item1",
"loopindex": 0,
"loopindex1": 1
})
cmd( 1):
echo """opening file: {{.loopitem}}"""
-
opening file: item1
-
.. ok
cmd( 2):
echo "hello"|grep "world"
-
-
.. failed(suppressed if it is not the last step)
task Finally:
---Step1: [
close_fileensure the opened file is closed
]
self: final context exec vars:
(*core.Cache)({
"loopitem": "item1",
"loopindex": 0,
"loopindex1": 1,
"up_runtime_task_layer_number": 2
})
cmd( 1):
echo "close the file ....."
-
close the file .....
-
.. ok
. ok
WARN: [Rescued in task level, but not advised!] - [setting rescue to yes/true to continue is not recommended
it is advised to locate root cause of the problem, fix it and re-run the task again
it is the best practice to test the execution in your ci pipeline to eliminate problems rather than dynamically fix using rescue]
===Task3: [task/sub_task_layer1/sub_task_layer2 ==> sub_task_layer2: without rescue, the execution will return a non-zero return code in shell and also report the error
with rescue, the program will return 0
]
----Step1: [step1: step 1 ]
self: final context exec vars:
(*core.Cache)({
"loopitem": "item2",
"loopindex": 1,
"loopindex1": 2,
"last_result": (*utils.ExecResult)({
Cmd: "echo \"close the file .....\"",
Code: 0,
Output: "close the file .....",
ErrMsg: ""
}),
"up_runtime_task_layer_number": 3
})
cmd( 1):
echo """opening file: {{.loopitem}}"""
-
opening file: item2
-
.. ok
cmd( 2):
echo "hello"|grep "world"
-
-
.. failed(suppressed if it is not the last step)
task Finally:
----Step1: [
close_fileensure the opened file is closed
]
self: final context exec vars:
(*core.Cache)({
"loopindex1": 2,
"up_runtime_task_layer_number": 3,
"loopitem": "item2",
"loopindex": 1
})
cmd( 1):
echo "close the file ....."
-
close the file .....
-
.. ok
. ok
WARN: [Rescued in task level, but not advised!] - [setting rescue to yes/true to continue is not recommended
it is advised to locate root cause of the problem, fix it and re-run the task again
it is the best practice to test the execution in your ci pipeline to eliminate problems rather than dynamically fix using rescue]
WARN: [Not rescued in task level] - [please assess the panic problem and cause, fix it before re-run the task]
task finally -> ERROR: Failed And Not Ignored! [You may want to continue and ignore the error]
Logs with different verbose level
Raw logs with different verbose level
Other references: