configure dcos 1.8 for centos 7.3

Configure DCOS 1.8 for Centos 7.3

configure-all-nodes.md

All Nodes

Install system requirements

https://dcos.io/docs/1.8/administration/installing/custom/system-requirements/

Install docker

https://dcos.io/docs/1.8/administration/installing/custom/system-requirements/install-docker-centos/

• you must have 2 cpu for all nodes or the navstar service will die... a lot
• you must have ntp enabled

Follow advanced install

https://dcos.io/docs/1.8/administration/installing/custom/advanced/

• the detect_ip script must be correct with #!/bin/bash at the top or ... fail
• also if you have multiple ipv4 addresses, it will fail

Make sure agent can connect by flushing agent configuration

if you need to restart the agent or change the agent configuration, make sure to clear its current config

rm -f /var/lib/mesos/slave/meta/slaves/latest
systemctl restart dcos-mesos-slave.service

set-static-ip-process.md

edit the eth0 device to assign a static ip also add a new UUID if cloned from previous image

vi /etc/sysconfig/network-scripts/ifcfg-eth0

I ended up struggling a bunch with NetworkManager and had to add a staticip.conf file to /etc/NetworkManager/conf.d/ with the following contents

[main]
no-auto-default=eth0

after that NeworkManager should leave your interface alone. You should remove any additional eth0 files from also remove any WiredConnection scripts dropped by NetworkManager

rm -rf /etc/sysconfig/network-scripts/eth0-0
rm -rf /etc/sysconfig/network-scripts/Wired_Connection_1

restart network.service

systemctl restart network.service

if network service complains, check again for duplicate interfaces in /etc/sysconfig/network-scripts/ you can also take the interfaces down

ifdown eth0
ifup eth0

check if the ip address is still assigned, if so flush the ip addresses and restart network manager again

ip addr flus dev eth0
systemctl restart network.service

show existing ip address to make sure only one is assigned

ip addr show eth0

try to ping google and the gateway

ping 192.168.1.1
ping google.com

if gateway ping works but google ping doesn't check /etc/resolv.conf or run dhclient

dhclient
ps aux | grep dhc
kill {pid of dhclient}  

NetworkManager may try to take over your /etc/resolv.conf, make sure to set the following in your /etc/sysconfig/network-scripts/ifcfg-eth0

PEERDNS=no

setup-ssh-keys.md

cd ~ mkdir .ssh cd .ssh

TODO: add key generation using ssh keygen

TODO: copy public key to authorized_keys

TODO: copy public key to authroized_keys2