-
Bug
-
Resolution: Fixed
-
Minor
-
jenkins: jenkins:2.123-alpine (running in docker)
Pipeline: Declarative: 1.2.9
Pipeline: Groovy: 2.53
When a timeout call is fired inside of a retry, retry is not being triggered and job execution is aborted, the only way to make it work is by surrounding the timeout operation with a try/catch.
without try/catch
Log output
Cancelling nested steps due to timeout
Execution result
Timeout has been exceeded Finished: ABORTED
with try/catch
Log output
Timeout set to expire after 2 sec without activity
Sleeping for 2 sec
Cancelling nested steps due to timeout
ERROR: catched timeout! org.jenkinsci.plugins.workflow.steps.FlowInterruptedException
Retrying
Execution result
Finished: SUCCESS
Examples to reproduce the issue
Failing example
node {
def timeoutSeconds = 3
stage('Preparation') { // for display purposes
retry(3){
timeout(activity: true, time: 2, unit: 'SECONDS') {
sleep(timeoutSeconds)
}
timeoutSeconds--
}
}
}
Working example
node {
def timeoutSeconds = 3
stage('Preparation') { // for display purposes
retry(3){
try{
timeout(activity: true, time: 2, unit: 'SECONDS') {
sleep(timeoutSeconds)
}
}catch(err){
timeoutSeconds--
script{
def user = err.getCauses()[0].getUser()
error "[${user}] catched timeout! $err"
}
}
}
}
}
- relates to
-
-
- Resolved
-
-
-
- Resolved
-
- links to
[JENKINS-51454] Pipeline retry operation doesn't retry when there is a timeout inside of it
Sam Van Oort
made changes -
| Component/s | New: pipeline-model-definition-plugin [ 21706 ] | |
| Component/s | Original: groovy-plugin [ 15549 ] | |
| Component/s | Original: pipeline [ 21692 ] |
Sam Van Oort
made changes -
| Assignee | Original: vjuranek [ vjuranek ] | New: Amita Bhasin [ ab ] |
Sam Van Oort
made changes -
| Assignee | Original: Amita Bhasin [ ab ] | New: Andrew Bayer [ abayer ] |
| Labels | Original: abort declarative pipeline retry timeout | New: abort declarative pipeline retry timeout triaged-2018-11 |
Luis Piedra-Márquez
added a comment - If capturing the `FlowInterruptedException` it will retry also when aborting the job for any other reason, like canceling. Definitively, timeout should thow a different exception.
Luis Piedra-Márquez
added a comment - If capturing the ` FlowInterruptedException` it will retry also when aborting the job for any other reason, like canceling. Definitively, timeout should thow a different exception. 
Basil Crow
added a comment - I can reproduce this error in a scripted pipeline as well. Basil Crow
made changes -
| Summary | Original: Declarative pipeline retry operation doesn't retry when there is a timeout inside of it | New: Pipeline retry operation doesn't retry when there is a timeout inside of it |
Basil Crow
made changes -
| Description |
Original:
When a declarative {{timeout}} call is fired inside of a {{retry}}, *retry is not being triggered* and *job execution is aborted*, the only way to make it work is by surrounding the {{timeout}} operation with a {{try/catch}}. h3. without try/catch Log output {code} Cancelling nested steps due to timeout {code} Execution result {code} Timeout has been exceeded Finished: ABORTED {code} h3. with try/catch Log output {code} Timeout set to expire after 2 sec without activity Sleeping for 2 sec Cancelling nested steps due to timeout ERROR: catched timeout! org.jenkinsci.plugins.workflow.steps.FlowInterruptedException Retrying {code} Execution result {code} Finished: SUCCESS {code} h2. Examples to reproduce the issue h3. Failing example {code:java} node { def timeoutSeconds = 3 stage('Preparation') { // for display purposes retry(3){ timeout(activity: true, time: 2, unit: 'SECONDS') { sleep(timeoutSeconds) } timeoutSeconds-- } } } {code} h3. Working example {code:java} node { def timeoutSeconds = 3 stage('Preparation') { // for display purposes retry(3){ try{ timeout(activity: true, time: 2, unit: 'SECONDS') { sleep(timeoutSeconds) } }catch(err){ timeoutSeconds-- script{ def user = err.getCauses()[0].getUser() error "[${user}] catched timeout! $err" } } } } } {code} |
New:
When a {{timeout}} call is fired inside of a {{retry}}, *retry is not being triggered* and *job execution is aborted*, the only way to make it work is by surrounding the {{timeout}} operation with a {{try/catch}}. h3. without try/catch Log output {code} Cancelling nested steps due to timeout {code} Execution result {code} Timeout has been exceeded Finished: ABORTED {code} h3. with try/catch Log output {code} Timeout set to expire after 2 sec without activity Sleeping for 2 sec Cancelling nested steps due to timeout ERROR: catched timeout! org.jenkinsci.plugins.workflow.steps.FlowInterruptedException Retrying {code} Execution result {code} Finished: SUCCESS {code} h2. Examples to reproduce the issue h3. Failing example {code:java} node { def timeoutSeconds = 3 stage('Preparation') { // for display purposes retry(3){ timeout(activity: true, time: 2, unit: 'SECONDS') { sleep(timeoutSeconds) } timeoutSeconds-- } } } {code} h3. Working example {code:java} node { def timeoutSeconds = 3 stage('Preparation') { // for display purposes retry(3){ try{ timeout(activity: true, time: 2, unit: 'SECONDS') { sleep(timeoutSeconds) } }catch(err){ timeoutSeconds-- script{ def user = err.getCauses()[0].getUser() error "[${user}] catched timeout! $err" } } } } } {code} |
It looks like the org.jenkinsci.plugins.workflow.steps.FlowInterruptedException exception could be the cause. If the timeout directive would throw another exception, would this solve the issue?