-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/49555/
-----------------------------------------------------------
(Updated Aug. 24, 2016, 5:22 p.m.)
Review request for mesos, Alexander Rukletsov, Benjamin Mahler, Gastón Kleiman,
Gilbert Song, Jie Yu, and Timothy Chen.
Changes
-------
Rebase.
Bugs: MESOS-5954
https://issues.apache.org/jira/browse/MESOS-5954
Repository: mesos
Description
-------
We change the mesos-executor to use health check via library way in
https://reviews.apache.org/r/49389/. To keep consistent, we replace
health check in mesos-docker-executor from binary way to library way.
In this patch, we rename `healthCheckDir` to `launcherDir` as well to
keep consistent with mesos-executor.
Diffs (updated)
-----
src/docker/executor.cpp 8d679cd33b6ddf3a5c11bb8c458a97b8809473ac
Diff: https://reviews.apache.org/r/49555/diff/
Testing
-------
Test with Marathon locally.
# 1. Success case.
I start `sleep 200` with the Docker container and use `ls /mnt/mesos/sandbox`
as the health check command.
According the qurey result of `http://localhost:8080/v2/apps/test-sleep`, could
see Marathon receive healthy status.
```
# truncate unnecessary messages
{
"app": {
"id": "/test-health",
"cmd": "sleep 200",
"container": {
"type": "DOCKER",
"volumes": [],
"docker": {
"image": "ubuntu",
"network": "HOST"
}
},
"healthChecks": [
{
"protocol": "COMMAND",
"command": {
"value": "ls /mnt/mesos/sandbox"
},
"gracePeriodSeconds": 300,
"intervalSeconds": 60,
"timeoutSeconds": 20,
"maxConsecutiveFailures": 3
}
],
"tasksStaged": 0,
"tasksRunning": 1,
"tasksHealthy": 1,
"tasksUnhealthy": 0,
"tasks": [
{
"id": "test-health.8d4ba884-585c-11e6-9fb5-0242e663b1f5",
"slaveId": "ec5359a2-9c63-4c2e-86da-3ed8ef94800b-S0",
"host": "127.0.0.1",
"startedAt": "2016-08-02T02:55:23.433Z",
"stagedAt": "2016-08-02T02:55:21.217Z",
"version": "2016-08-02T02:48:31.054Z",
"ipAddresses": [
{
"ipAddress": "127.0.0.1",
"protocol": "IPv4"
}
],
"appId": "/test-health",
"healthCheckResults": [
{
"alive": true,
"consecutiveFailures": 0,
"firstSuccess": "2016-08-02T02:55:23.652Z",
"lastFailure": null,
"lastSuccess": "2016-08-02T02:55:23.652Z",
"lastFailureCause": null,
"taskId": "test-health.8d4ba884-585c-11e6-9fb5-0242e663b1f5"
}
]
}
]
}
}
```
# 2. Failed case.
Then I start a similar application which use `false` as health check command,
could see Marathon receive unhealthy status.
```
# truncate unnecessary messages
{
"app": {
"id": "/test-unhealth",
"cmd": "sleep 200",
"container": {
"type": "DOCKER",
"volumes": [],
"docker": {
"image": "ubuntu",
"network": "HOST"
}
},
"healthChecks": [
{
"protocol": "COMMAND",
"command": {
"value": "false"
},
"gracePeriodSeconds": 5,
"intervalSeconds": 3,
"timeoutSeconds": 20,
"maxConsecutiveFailures": 3
}
],
"tasksStaged": 0,
"tasksRunning": 1,
"tasksHealthy": 0,
"tasksUnhealthy": 1,
"tasks": [
{
"id": "test-unhealth.48a6318a-585e-11e6-9fb5-0242e663b1f5",
"slaveId": "ec5359a2-9c63-4c2e-86da-3ed8ef94800b-S0",
"host": "127.0.0.1",
"startedAt": "2016-08-02T03:07:47.310Z",
"stagedAt": "2016-08-02T03:07:45.041Z",
"ipAddresses": [
{
"ipAddress": "127.0.0.1",
"protocol": "IPv4"
}
],
"appId": "/test-unhealth",
"healthCheckResults": [
{
"alive": false,
"consecutiveFailures": 1,
"firstSuccess": null,
"lastFailure": "2016-08-02T03:07:53.800Z",
"lastSuccess": null,
"lastFailureCause": "",
"taskId": "test-unhealth.48a6318a-585e-11e6-9fb5-0242e663b1f5"
}
]
}
]
}
}
```
Thanks,
haosdent huang
