[
https://issues.apache.org/jira/browse/MESOS-1541?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14046411#comment-14046411
]
Benjamin Mahler commented on MESOS-1541:
----------------------------------------
In the logs you posted above, it looks as though the slave is able to create a
connection with the master but the master is unable to create a connection back
to the slave.
OK: 11.11.42.73 → 12.12.249.207
BAD: 12.12.249.207 → 11.11.42.73
Can you confirm whether 12.12.249.207 can create a connection with 11.11.42.73?
> Mesos slave continuous disconnection
> -------------------------------------
>
> Key: MESOS-1541
> URL: https://issues.apache.org/jira/browse/MESOS-1541
> Project: Mesos
> Issue Type: Bug
> Components: master
> Affects Versions: 0.19.0
> Environment: Oracle Enterprise Linux
> Reporter: Umesh Batra
> Priority: Blocker
> Labels: disconnection, master, mesos, offer, rejection, slave
>
> I am seeing continuous disconnections and offer rejections in master's logs,
> its happening almost every 10-20 seconds
>
> As per suggestions from various reads
>
> I tried two approaches,
>
> 1. Setting --ip, --hostname and --port flags on the master and slave
> processes
>
> Here's one of the master/slave process details (ps -ef output)
>
> master:
>
> /usr/local/sbin/mesos-master --ip=11.11.42.73
> --hostname=mesosd-lapp01.mylab.com --port=5050 --work_dir=/mesos/data/master
> --log_dir=/mesos/logs/master --quorum=1
> --zk=zk://11.11.42.73:2181,11.11.42.78:2181,11.11.42.79:2181/mesos
>
> slave:
>
> /usr/local/sbin/mesos-slave --ip=12.12.249.207 --hostname=pod1-08.mylab.com
> --port 5051 --work_dir=/mesos/data/slave --log_dir=/mesos/logs/slave
> --master=zk://11.11.42.73:2181,11.11.42.78:2181,11.11.42.79:2181/mesos
> --isolation=cgroups/cpu,cgroups/mem --cgroups_hierarchy=/cgroup
> --cgroups_root=/cgroup
>
> 2. I also tried specifying external IP and real hostname in /etc/hosts file
> e.g.
>
> 10.11.42.73 mesosd-lapp01.mylab.com in master host and
> 12.12.249.207 pod1-08.mylab.com in slave host
>
> appreciate your timely response.
>
> - rgds,
> Umesh
>
>
--
This message was sent by Atlassian JIRA
(v6.2#6252)
