Details
-
Type:
Bug
-
Status: Resolved
-
Priority:
Low
-
Resolution: Won't Do
-
Affects Version/s: None
-
Fix Version/s: None
-
Component/s: marathon-deployments
-
Labels:
-
Epic Link:
-
Story Points:0
Description
Marathon Version: 1.3.6
Step to reproduce:
1. start a simple web server app with health check, with 1 instance, wait unit running
{
"id": "/test",
"cmd": null,
"cpus": 1,
"mem": 128,
"instances": 1,
"container": {
"type": "DOCKER",
"volumes": [],
"docker": {
"image": "httpd:2.4",
"network": "BRIDGE",
"portMappings": [
{
"containerPort": 80,
"protocol": "tcp"
}
]
}
},
"healthChecks": [
{
"path": "/",
"protocol": "HTTP",
"portIndex": 0,
"gracePeriodSeconds": 60,
"intervalSeconds": 20,
"timeoutSeconds": 20,
"maxConsecutiveFailures": 3,
"ignoreHttp1xx": false
}
]
}
2. modify app configuration, change app cmd to "sleep 20", and upgrade
Expected result:
1. Newly launched tasks fail health check
2. Newly launched task finish
3. Marathon launch another task after the task finish
4. repeat step 1
Actual result:
1. Newly launched tasks fail health check
2. Newly launched task finish
3. From now on, marathon deployment get stuck and do not try to launch new task
I think this affects marathon 1.3.x and 1.1.x
