You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mesos.apache.org by "James Peach (JIRA)" <ji...@apache.org> on 2018/01/26 18:50:02 UTC
[jira] [Created] (MESOS-8496) Support revocable resources for task
groups in mesos-execute.
James Peach created MESOS-8496:
----------------------------------
Summary: Support revocable resources for task groups in mesos-execute.
Key: MESOS-8496
URL: https://issues.apache.org/jira/browse/MESOS-8496
Project: Mesos
Issue Type: Bug
Components: oversubscription
Reporter: James Peach
When you use {{mesos-execute}} to launch a task group that used revocable resources, it tries to accept an invalid combination of revocable and non-revocable resources:
With the following task info:
{noformat}
{
"tasks": [
{
"name": "run-sleep-task-group-0",
"task_id": {
"value": "71499a12-5999-4417-beea-82bdc4a840f8"
},
"agent_id": {
"value": ""
},
"resources": [
{
"name": "cpus",
"type": "SCALAR",
"revocable": {},
"scalar": {
"value": 1
}
},
{
"name": "mem",
"type": "SCALAR",
"scalar": {
"value": 32
}
},
{
"name": "disk",
"type": "SCALAR",
"scalar": {
"value": 16
}
}
],
"command": {
"value": "sleep 10000"
},
}
]
}
{noformat}
You get this problem:
{noformat}I0126 10:32:32.701838 9275 logging.cpp:208] Logging to STDERR
I0126 10:32:32.709440 9275 process.cpp:1218] libprocess is initialized on 17.228.224.108:45099 with 12 worker threads
I0126 10:32:32.710484 9297 scheduler.cpp:188] Version: 1.6.0
I0126 10:32:32.718277 9290 scheduler.cpp:311] Using default 'basic' HTTP authenticatee
I0126 10:32:32.718789 9296 scheduler.cpp:494] New master detected at master@17.228.224.108:5050
I0126 10:32:32.720448 9296 scheduler.cpp:503] Waiting for 422.561438ms before initiating a re-(connection) attempt with the master
I0126 10:32:33.146246 9290 scheduler.cpp:385] Connected with the master at http://17.228.224.108:5050/master/api/v1/scheduler
I0126 10:32:33.147275 9286 scheduler.cpp:247] Sending SUBSCRIBE call to http://17.228.224.108:5050/master/api/v1/scheduler
I0126 10:32:33.158856 9290 scheduler.cpp:739] Enqueuing event SUBSCRIBED received from http://17.228.224.108:5050/master/api/v1/scheduler
I0126 10:32:33.159373 9290 scheduler.cpp:739] Enqueuing event HEARTBEAT received from http://17.228.224.108:5050/master/api/v1/scheduler
Subscribed with ID bed65d63-f377-4dda-bad9-9ced9ff37cf8-0004
I0126 10:32:33.161146 9288 scheduler.cpp:739] Enqueuing event OFFERS received from http://17.228.224.108:5050/master/api/v1/scheduler
Submitted task group with tasks [ d65ea022-4c76-4664-a3a1-b5f302d9b935 ] to agent '0afcc784-79ce-428b-9a75-2fce66a20e23-S1'
I0126 10:32:33.163039 9290 scheduler.cpp:247] Sending ACCEPT call to http://17.228.224.108:5050/master/api/v1/scheduler
I0126 10:32:33.170541 9288 scheduler.cpp:739] Enqueuing event UPDATE received from http://17.228.224.108:5050/master/api/v1/scheduler
Received status update TASK_ERROR for task 'd65ea022-4c76-4664-a3a1-b5f302d9b935'
message: 'Task group and executor mix revocable and non-revocable resources: Cannot use both revocable and non-revocable 'cpus' at the same time'
source: SOURCE_MASTER
reason: REASON_TASK_GROUP_INVALID
{noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)