You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@mesos.apache.org by Nils-Helge Garli Hegvik <ni...@gmail.com> on 2015/02/12 22:01:12 UTC
Troubles with setting up master/slave
Hi!
I'm following [1] to set up a master/slave configuration with Vagrant.
I have set up networking and hostnames so the master and slave can
resolve each other. I can get to the ui, and see that a framework
(marathon) and a slave is registered. I can also successfully verify
the installation by issuing a task with
mesos-execute --master=$MASTER --name="cluster-test" --command="sleep 5"
on the master. However, when launching the same task from the slave
vm, it fails. It seems to go into an error loop, where the master
keeps logging
---
I0212 20:28:36.912345 9244 master.cpp:3843] Sending 1 offers to
framework 20150212-200948-46704832-5050-9196-0000 (marathon) at
scheduler-9cf62fe9-c24d-451e-821d-f3e9aefa8d2c@127.0.1.1:40413
I0212 20:28:36.984591 9242 master.cpp:2344] Processing reply for
offers: [ 20150212-200948-46704832-5050-9196-O175 ] on slave
20150212-200948-46704832-5050-9196-S0 at slave(1)@192.168.200.3:5051
(192.168.200.3) for framework 20150212-200948-46704832-5050-9196-0000
(marathon) at scheduler-9cf62fe9-c24d-451e-821d-f3e9aefa8d2c@127.0.1.1:40413
I0212 20:28:36.985373 9242 hierarchical_allocator_process.hpp:563]
Recovered cpus(*):2; mem(*):496; disk(*):32297; ports(*):[31000-32000]
(total allocatable: cpus(*):2; mem(*):496; disk(*):32297;
ports(*):[31000-32000]) on slave 20150212-200948-46704832-5050-9196-S0
from framework 20150212-200948-46704832-5050-9196-0000
I0212 20:28:37.749330 9242 http.cpp:478] HTTP request for '/master/state.json'
I0212 20:28:37.820500 9242 master.cpp:1383] Received registration
request for framework '' at
scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
I0212 20:28:37.821678 9242 master.cpp:1447] Registering framework
20150212-200948-46704832-5050-9196-0059 () at
scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
I0212 20:28:37.822754 9238 hierarchical_allocator_process.hpp:329]
Added framework 20150212-200948-46704832-5050-9196-0059
I0212 20:28:37.823096 9244 master.cpp:789] Framework
20150212-200948-46704832-5050-9196-0059 () at
scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
disconnected
I0212 20:28:37.824023 9244 master.cpp:1752] Disconnecting framework
20150212-200948-46704832-5050-9196-0059 () at
scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
I0212 20:28:37.824358 9244 master.cpp:1768] Deactivating framework
20150212-200948-46704832-5050-9196-0059 () at
scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
I0212 20:28:37.824725 9244 master.cpp:811] Giving framework
20150212-200948-46704832-5050-9196-0059 () at
scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664 0ns to
failover
W0212 20:28:37.825021 9244 master.cpp:3726] Master returning
resources offered to framework 20150212-200948-46704832-5050-9196-0059
because the framework has terminated or is inactive
I0212 20:28:37.825500 9244 hierarchical_allocator_process.hpp:405]
Deactivated framework 20150212-200948-46704832-5050-9196-0059
---
I do notice that the ip of the scheduler reports as 127.0.1.1:47664,
but I have set all of /etc/mesos-master/hostname, /etc/mesos-master/ip
and /etc/marathon/conf/hostname to the private network of the vagrant
vms.
[1] - http://mesosphere.com/docs/getting-started/datacenter/install/
Re: Troubles with setting up master/slave
Posted by Adam Bordelon <ad...@mesosphere.io>.
You could also try setting the LIBPROCESS_IP environment variable, which
should override the IP that the scheduler reports for its PID.
On Thu, Feb 12, 2015 at 1:01 PM, Nils-Helge Garli Hegvik <ni...@gmail.com>
wrote:
> Hi!
>
> I'm following [1] to set up a master/slave configuration with Vagrant.
> I have set up networking and hostnames so the master and slave can
> resolve each other. I can get to the ui, and see that a framework
> (marathon) and a slave is registered. I can also successfully verify
> the installation by issuing a task with
>
> mesos-execute --master=$MASTER --name="cluster-test" --command="sleep 5"
>
> on the master. However, when launching the same task from the slave
> vm, it fails. It seems to go into an error loop, where the master
> keeps logging
>
> ---
>
> I0212 20:28:36.912345 9244 master.cpp:3843] Sending 1 offers to
> framework 20150212-200948-46704832-5050-9196-0000 (marathon) at
> scheduler-9cf62fe9-c24d-451e-821d-f3e9aefa8d2c@127.0.1.1:40413
>
> I0212 20:28:36.984591 9242 master.cpp:2344] Processing reply for
> offers: [ 20150212-200948-46704832-5050-9196-O175 ] on slave
> 20150212-200948-46704832-5050-9196-S0 at slave(1)@192.168.200.3:5051
> (192.168.200.3) for framework 20150212-200948-46704832-5050-9196-0000
> (marathon) at
> scheduler-9cf62fe9-c24d-451e-821d-f3e9aefa8d2c@127.0.1.1:40413
>
> I0212 20:28:36.985373 9242 hierarchical_allocator_process.hpp:563]
> Recovered cpus(*):2; mem(*):496; disk(*):32297; ports(*):[31000-32000]
> (total allocatable: cpus(*):2; mem(*):496; disk(*):32297;
> ports(*):[31000-32000]) on slave 20150212-200948-46704832-5050-9196-S0
> from framework 20150212-200948-46704832-5050-9196-0000
>
> I0212 20:28:37.749330 9242 http.cpp:478] HTTP request for
> '/master/state.json'
>
> I0212 20:28:37.820500 9242 master.cpp:1383] Received registration
> request for framework '' at
> scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
>
> I0212 20:28:37.821678 9242 master.cpp:1447] Registering framework
> 20150212-200948-46704832-5050-9196-0059 () at
> scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
>
> I0212 20:28:37.822754 9238 hierarchical_allocator_process.hpp:329]
> Added framework 20150212-200948-46704832-5050-9196-0059
>
> I0212 20:28:37.823096 9244 master.cpp:789] Framework
> 20150212-200948-46704832-5050-9196-0059 () at
> scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
> disconnected
>
> I0212 20:28:37.824023 9244 master.cpp:1752] Disconnecting framework
> 20150212-200948-46704832-5050-9196-0059 () at
> scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
>
> I0212 20:28:37.824358 9244 master.cpp:1768] Deactivating framework
> 20150212-200948-46704832-5050-9196-0059 () at
> scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664
>
> I0212 20:28:37.824725 9244 master.cpp:811] Giving framework
> 20150212-200948-46704832-5050-9196-0059 () at
> scheduler-f5510035-c6ea-4074-ad49-2b16d855b127@127.0.1.1:47664 0ns to
> failover
>
> W0212 20:28:37.825021 9244 master.cpp:3726] Master returning
> resources offered to framework 20150212-200948-46704832-5050-9196-0059
> because the framework has terminated or is inactive
>
> I0212 20:28:37.825500 9244 hierarchical_allocator_process.hpp:405]
> Deactivated framework 20150212-200948-46704832-5050-9196-0059
>
> ---
>
> I do notice that the ip of the scheduler reports as 127.0.1.1:47664,
> but I have set all of /etc/mesos-master/hostname, /etc/mesos-master/ip
> and /etc/marathon/conf/hostname to the private network of the vagrant
> vms.
>
>
> [1] - http://mesosphere.com/docs/getting-started/datacenter/install/
>