goodrain/rainbond-install

Saltstack recipe to deploy Rainbond (abandoned)

Status(Abandoned)

Install Rainbond that version after 5.0.0, Please reference https://github.com/goodrain/rainbond-ansible

Description

A collection of Salt states used to provision an application-centric platform Rainbond.

Support for:

• Rainbond Manage node - single node with Mysql database
• Rainbond Compute node - it is not automatically added to Rainbond environment
• Rainbond environment support - support Kubernetes
• Docker registry - automatically provisioned on Manage start
• Support for different provision

Plugins Version:

• Distribution
• Kubernetes v1.6.4
• Docker v1.12.6
• Calico v2.4.1
• Etcd v3.2.13
• Rainbond Core Plugins v3.7.2
• Rainbond UI v3.7.2
• Salt v2018.3.2

Operations Guide

Prerequisites

• Distribution-recommendation
• Custom-configuration
• Software-and-Hardware-Requirements

Install

• Quick start

# Stable v3.7.2(v3.7)

git clone --depth 1 -b v3.7 https://github.com/goodrain/rainbond-install.git /opt/rainbond/install
cd /opt/rainbond/install
./setup.sh

or

grctl init 

# previous v3.6.1

git clone --depth 1 -b v3.6 https://github.com/goodrain/rainbond-install.git
cd rainbond-install
./setup.sh

# previous v3.5.2
git clone --depth 1 -b v3.5 https://github.com/goodrain/rainbond-install.git
cd rainbond-install
./setup.sh

Documents

You can find all the documentation in the troubleshooting install/Docs.

License

This project is licensed under the Apache License, Version 2.0. See LICENSE for the full license text.

rainbond-install's Issues

install salt-* error

bootstrap-salt.sh
Failed to run install_centos_stable()!!!

how to generate host_uuid

web端无法访问容器

support traefik

#61

Rainbond integrated customized docker, Please uninstall it first.

--- Check system environment... ---
FAILED Rainbond integrated customized docker, Please uninstall it first.

Quit rainbond install program.

rename system_info.sls to goodrain.sls

use /etc/hosts resolving rainbond internal domain

• Don't modify user system resolv.conf config
• /etc/hosts binding internal domain

Adjust domain generation strategy

• Custom domain name
• Use Alidns

update etcd-proxy

• update etcd-proxy start script

Generate domain error.

centos 7.4

使用安装脚本，在make_domain步骤出错

在分配泛解析域名这一步报错退出
master分支和dev分支均存在此问题，请问如何解决？

salt 错误

start init
Could not access /var/cache/salt/master. Path does not exist.
start install storage
Could not access /var/cache/salt/master. Path does not exist.
start install docker
Could not access /var/cache/salt/master. Path does not exist.
start etcd
Could not access /var/cache/salt/master. Path does not exist.
start network plugin calico
Could not access /var/cache/salt/master. Path does not exist.
start k8s server
Could not access /var/cache/salt/master. Path does not exist.
start node
Could not access /var/cache/salt/master. Path does not exist.
start database mysql
Could not access /var/cache/salt/master. Path does not exist.
start plugins
Could not access /var/cache/salt/master. Path does not exist.
Could not access /var/cache/salt/master. Path does not exist.
start proxy
Could not access /var/cache/salt/master. Path does not exist.
start prom
Could not access /var/cache/salt/master. Path does not exist.
--- will install compute node. ---
start kubelet
Could not access /var/cache/salt/master. Path does not exist.

安装前需要检测端口是否被占用

53
80
443
2379
2380
3306
4001
6060
6100
6443
7070
8181
9999

salt-ssh 安装

支持修改域名解析

update pillar inet-ip

废弃inet-ip 配置，使用manage ip

# old
inet-ip: 172.16.0.1
# new
manage:
    members:
    - host: 172.16.0.1
      name: manage01
    - host: 172.16.0.2
      name: manage02
    - host: 172.16.0.3
      name: manage03

centos 7.4 make_domain 步骤卡住

• 去掉make_domain step
• 使用api接口获取domain

#72 #73

docker registry mirrors support customization

支持修改docker镜像加速配置，提升镜像拉取速度，目前默认使用docker-cn配置

rm docker-patch.sh

修改容器ulimit，需要修改的值

修改/usr/lib/systemd/system/docker.service
LimitNOFILE=102400
LimitNPROC=102400

默认清除防火墙规则

config rsync

检查端口是否被占用有问题

udp6       0      0 :::19532                :::*                                2266/dhclient

first master node install salt by salt-ssh

region_info数据重复

数据库中关于region_info数据重复

script install err

centos 7.0

Start install db ...
manage01:
Name: docker pull rainbond/rbd-db:3.5 - Function: cmd.run - Result: Changed Started: - 23:39:41.327276 Duration: 136468.565 ms

      ID: db-upstart
Function: cmd.run
    Name: dc-compose up -d rbd-db
  Result: False
 Comment: Command "dc-compose up -d rbd-db" run
 Started: 23:41:57.796429
Duration: 280.822 ms
 Changes:   
          ----------
          pid:
              4526
          retcode:
              127
          stderr:
              /usr/local/bin/docker-compose: error while loading shared libraries: libz.so.1: failed to map segment from shared object: Operation not permitted
          stdout:

Summary for manage01

Succeeded: 1 (changed=2)
Failed: 1

Total states run: 2
Total run time: 136.749 s
ERROR: Minions returned with non-zero exit code
Quit rainbond install program.

Unable to access public network ip

cat /srv/pillar/goodrain.sls
inet-ip: 10.66.176.107
public-ip:

Expand a manage node

retrieves public/private ip of a node

salt-ssh/salt manage01 network.ip_addrs type=public
salt-ssh/salt manage01 network.ip_addrs type=private

将主机uuid与ip对应信息写入到 /etc/hosts 文件中

entrance missing configuration file

entrance configuration file: host_uuid.conf

Remove rbd-proxy

update goodrain.me to lb

install.sh 报错

./install.sh: line 49: [: ==: unary operator expected
./install.sh: line 52: [: ==: unary operator expected

system: centos 7.4

国外源每次都超时，估计这个脚本只能国外机器用...

lb缺少forward文件

拉取镜像失败

          ID: runner-pull
    Function: cmd.run
        Name: docker pull rainbond/runner; docker tag rainbond/runner goodrain.me/runner;docker push goodrain.me/runner
      Result: False
     Comment: Command "docker pull rainbond/runner; docker tag rainbond/runner goodrain.me/runner;docker push goodrain.me/runner" run
     Started: 22:21:03.276421
    Duration: 211625.232 ms
     Changes:
              ----------
              pid:
                  26290
              retcode:
                  1
              stderr:
                  Put http://goodrain.me/v1/repositories/runner/: dial tcp 172.30.42.1:80: getsockopt: connection refused

fix rbd-repo restart

install Failed salt

--- Installing salt ... ---
FAILED Failed to install salt!

安装时会删除系统文件

/root/rainbond-install/setup.sh 
[ ! -d $PILLAR_DIR ] && mkdir $PILLAR_DIR || rm $PILLAR_DIR/* -rf

./scripts/check.sh: line 277

--- Installing salt ... ---
./scripts/check.sh: line 277: Echo_Error: command not found
--- There seems to be some problem here, You can through the error log(./logs/error.log) to get the detail information ---

rbd-db exit

Generate the ssh key

should not overwrite the original key.

node uuid和kubelet uuid不一致

hostIP={{ grains['fqdn_ip4'] }}
hostIP=['192.168.199.13']

网卡配置中如果包含dns配置，启动时会覆盖 /etc/resolv.conf 配置

需要安装时检查，或删除网卡配置文件中的dns配置信息，或者将云帮集群的dns配置，写到网卡配置文件中。

prometheus重启

目录权限不对

pillar.domain is not defined

配置host_id_list.conf

支持多个host_uuid

Expand node for rainbond

上线节点提示node is not k8s node or it not down

[root@manage01 ~]# grctl node list
+--------------------------------------+--------------+----------+----------+----------+--------+-------------+-------+
| Uid                                  | IP           | HostName | NodeRole | NodeMode | Alived | Schedulable | Ready |
+--------------------------------------+--------------+----------+----------+----------+--------+-------------+-------+
| 8e047f77-1b0f-4cb8-88b5-0f3ae14d89b5 | 172.16.0.168 | manage01 | manage   | master   | true   | N/A         | false |
+--------------------------------------+--------------+----------+----------+----------+--------+-------------+-------+
grctl node up 8e047f77-1b0f-4cb8-88b5-0f3ae14d89b5
400:error with code 400 , {"msg":"node is not k8s node or it not down"}

mysql db init failure

After mysql install, sleep 15 but db could not start. Whether to loop check port 3306 or sock file /data/tmp/mysqld.sock?