d2iq-archive / marathon

Deploy and manage containers (including Docker) on top of Apache Mesos at scale.
https://mesosphere.github.io/marathon/
Apache License 2.0
4.07k stars 843 forks source link

no matching offer for /group/app #1451

Closed lins05 closed 9 years ago

lins05 commented 9 years ago

System: ubuntu 14.04 LTS 64bit Marathon Version: 0.8.1-1.0.171.ubuntu1404 Mesos Version: 0.21.1-1.1.ubuntu1404

An application is not deployed even though the cluster has enough resource. See the logs below.

May  4 11:19:31 head2 marathon[27492]: [2015-05-04 11:19:31,443] INFO No matching offer for /xxx-group/web (need cpus=0.5, mem=384.0,
 disk=0.0, ports=Vector(0)) : id {
May  4 11:19:31 head2 marathon[27492]:   value: "20150428-051346-1388575636-5050-26039-O833496"
May  4 11:19:31 head2 marathon[27492]: }
May  4 11:19:31 head2 marathon[27492]: framework_id {
May  4 11:19:31 head2 marathon[27492]:   value: "20141126-052351-1405352852-5050-32174-0000"
May  4 11:19:31 head2 marathon[27492]: }
May  4 11:19:31 head2 marathon[27492]: slave_id {
May  4 11:19:31 head2 marathon[27492]:   value: "20150419-123325-1405352852-5050-1561-S11"
May  4 11:19:31 head2 marathon[27492]: }
May  4 11:19:31 head2 marathon[27492]: hostname: "mesos-slave.example.com"
May  4 11:19:31 head2 marathon[27492]: resources {
May  4 11:19:31 head2 marathon[27492]:   name: "cpus"
May  4 11:19:31 head2 marathon[27492]:   type: SCALAR
May  4 11:19:31 head2 marathon[27492]:   scalar {
May  4 11:19:31 head2 marathon[27492]:     value: 0.5
May  4 11:19:31 head2 marathon[27492]:   }
May  4 11:19:31 head2 marathon[27492]:   role: "*"
May  4 11:19:31 head2 marathon[27492]: }
May  4 11:19:31 head2 marathon[27492]: resources {
May  4 11:19:31 head2 marathon[27492]:   name: "mem"
May  4 11:19:31 head2 marathon[27492]:   type: SCALAR
May  4 11:19:31 head2 marathon[27492]:   scalar {
May  4 11:19:31 head2 marathon[27492]:     value: 384.0
May  4 11:19:31 head2 marathon[27492]:   }
May  4 11:19:31 head2 marathon[27492]:   role: "*"
May  4 11:19:31 head2 marathon[27492]: }
May  4 11:19:31 head2 marathon[27492]: resources {
May  4 11:19:31 head2 marathon[27492]:   name: "ports"
May  4 11:19:31 head2 marathon[27492]:   type: RANGES
May  4 11:19:31 head2 marathon[27492]:   ranges {
May  4 11:19:31 head2 marathon[27492]:     range {
May  4 11:19:31 head2 marathon[27492]:       begin: 31178
May  4 11:19:31 head2 marathon[27492]:       end: 31178
May  4 11:19:31 head2 marathon[27492]:     }
May  4 11:19:31 head2 marathon[27492]:   }
May  4 11:19:31 head2 marathon[27492]:   role: "*"

For me the offer seems enough for the app, why does marathon consider it "no matching"? Or did I get it wrong somewhere? There were also several other similar offers being considered "no matching" in the log.

kolloch commented 9 years ago

Is there any useful information in the log before your log excerpt that indicate what resources/constraints aren't satisfied?

If you specify just slightly less resources, does it work then? Not that it is a rounding issue or the like.

lins05 commented 9 years ago

Is there any useful information in the log before your log excerpt that indicate what resources/constraints aren't satisfied?

No, I don't find any.

If you specify just slightly less resources, does it work then? Not that it is a rounding issue or the like.

There were other offers that's more than engough, e.g.

May  4 11:19:27 head2 marathon[27492]: [2015-05-04 11:19:27,441] INFO No matching offer for /group/app (need cpus=0.5, mem=384.0, disk=0.0, ports=Vector(0)) : id {
May  4 11:19:27 head2 marathon[27492]:   value: "20150428-051346-1388575636-5050-26039-O833490"
May  4 11:19:27 head2 marathon[27492]: }
May  4 11:19:27 head2 marathon[27492]: framework_id {
May  4 11:19:27 head2 marathon[27492]:   value: "20141126-052351-1405352852-5050-32174-0000"
May  4 11:19:27 head2 marathon[27492]: }
May  4 11:19:27 head2 marathon[27492]: slave_id {
May  4 11:19:27 head2 marathon[27492]:   value: "20150414-203931-1388575636-5050-26585-S14"
May  4 11:19:27 head2 marathon[27492]: }
May  4 11:19:27 head2 marathon[27492]: hostname: "some-slave.example.com"
May  4 11:19:27 head2 marathon[27492]: resources {
May  4 11:19:27 head2 marathon[27492]:   name: "cpus"
May  4 11:19:27 head2 marathon[27492]:   type: SCALAR
May  4 11:19:27 head2 marathon[27492]:   scalar {
May  4 11:19:27 head2 marathon[27492]:     value: 1.25
May  4 11:19:27 head2 marathon[27492]:   }
May  4 11:19:27 head2 marathon[27492]:   role: "*"
May  4 11:19:27 head2 marathon[27492]: }
May  4 11:19:27 head2 marathon[27492]: resources {
May  4 11:19:27 head2 marathon[27492]:   name: "mem"
May  4 11:19:27 head2 marathon[27492]:   type: SCALAR
May  4 11:19:27 head2 marathon[27492]:   scalar {
May  4 11:19:27 head2 marathon[27492]:     value: 6400.0
May  4 11:19:27 head2 marathon[27492]:   }
May  4 11:19:27 head2 marathon[27492]:   role: "*"
May  4 11:19:27 head2 marathon[27492]: }

So I think it's not because of rounding errors.

aquamatthias commented 9 years ago

Hey @lins05 Can you provide the configuration of the mater and the slave as well as the definition of the application? Apparently it is not possible to say something on the logs that are provided.

kolloch commented 9 years ago

We cannot proceed without more information...