The heartbeat of High availability cluster resource management based on CRM (II.)

The heartbeat of high-availability cluster resource management based on CRM

1, the working model of the cluster:

a/p: Two nodes, working with master preparation model

N-m n>m,n nodes, M services

N-N:N nodes, N services

A/A: Dual Master Model:

2. How to transfer resources

Rgmanager:failover Domain Priority

Pacemaker

Resource stickiness:

Resource constraints (three types):

Location constraints: Resources tend to be on that node

INF: Infinity

N:

-N:

-inf: Negative Infinity

Permutation constraints: the tendency of resources to run on the same node

Inf:

-inf:

Order constraints: The starting and closing order of resources

3. How to make three resources in Web service: VIP, httpd, and filesystem running on the same node

1. Arranging constraints

2. Resource groups (resource group)

4. How to handle resources running on the current node if the node is not a member of a cluster node

Stopped: Stop

Ignore: Ignore

Freeze: Do not connect to new requests

Suicide: Kill the server

5, a resource just when the configuration is complete, whether to start

Target-role?

6. RA type

Heartbeat legacy

Lsb

OCF

STONITH

7. Resource type

Primitive,native: Master resource, can only run on one node

Group: Groups Resources

Clone: Cloning A resource

Total clones, maximum number of clones to run per node

Stonith cluster filesystem

Master/salve: Master-Slave resources

8. Distributed Lock:

/usr/lib64/heartbeat

hearsources2cib.py

9. Graphical Configuration

Ha.cf

CRM on

/usr/lib64/heartbeat/ha_propagate transferring a configuration file to another node

10. Installing the GUI

Heartbeat v2 using CRM as Ijiqun Explorer: Need to add in HA.CF

CRM on

CRM via MGMTD Integrated Monitoring 5560/tcp

The host that needs to start Hb_gui adds a password to the Hacluster user and starts with Hb_gui

With quorum: Statutory number of votes

Without quorum: no legal votes

11. Define a highly available Web service

VIP

httpd

from

to: based on it

Web Service

Vip

httpd

Nfs

Note Haresources is not compatible with CRM and is not read by CRM

Second, the configuration

1, HA.CF

[Email protected] heartbeat]# VIM/ETC/HA.D/HA.CF

Mcast eth0 225.0.100.19 694 1 0

CRM on

[Email protected] heartbeat]#/usr/lib64/heartbeat/ha_propagate

Propagating HA configuration files to node datanode4.abc.com.

HA.CF 100% 10KB 10.4kb/s 00:00

Authkeys 100% 694 0.7kb/s 00:00

Setting HA startup configuration on Node datanode4.abc.com.

2. Note that haresources is not compatible with CRM and is not read by CRM

[Email protected] heartbeat]# Mv/etc/ha.d/haresources/root

The bottom MV is Datanode4 's mainframe.

[Email protected] ha.d]# MV haresources/root/

[[Email protected] heartbeat]# service heartbeat start

LOGD is already running

Starting High-availability Services:

Done.

[[email protected] heartbeat]# ssh datanode4 ' service heartbeat start '

LOGD is already running

Starting High-availability Services:

Done.

3. View Logs

[Email protected] heartbeat]# tail-f/var/log/messages

June 16:00:29 SNN CRMD: [2223]: notice:populate_cib_nodes:Node:datanode4.abc.com (uuid:0862d824-047e-4826-9e26-21a7 603F53C8)

June 16:00:30 SNN CRMD: [2223]: notice:populate_cib_nodes:Node:snn.abc.com (uuid:6009ca6a-56eb-4d35-872e-3b8dc0fc98 51)

June 16:00:30 SNN CRMD: [2223]: info:do_ha_control:Connected to Heartbeat

June 16:00:30 SNN CRMD: [2223]: Info:do_ccm_control:CCM connection established ... waiting for first callback

June 16:00:30 SNN CRMD: [2223]: info:do_started:Delaying start, CCM (0000000000100000) not connected

June 16:00:30 SNN CRMD: [2223]: info:crmd_init:Starting Crmd ' s Mainloop

June 16:00:30 SNN CRMD: [2223]: Notice:crmd_client_status_callback:Status update:client SNN.ABC.COM/CRMD now have stat US [Online]

June 16:00:30 SNN CRMD: [2223]: Notice:crmd_client_status_callback:Status update:client SNN.ABC.COM/CRMD now have stat US [Online]

June 16:00:30 SNN CRMD: [2223]: notice:crmd_client_status_callback:Status update:client Datanode4.abc.com/crmd now ha s status [online]

June 16:00:30 SNN CIB: [2219]: Info:mem_handle_event:Got an event oc_ev_ms_new_membership from CCM

June 16:00:30 SNN CIB: [2219]: info:mem_handle_event:instance=5, nodes=2, new=2, lost=0, n_idx=0, new_idx=0, old_idx=4

June 16:00:30 SNN CIB: [2219]: info:cib_ccm_msg_callback:PEER:datanode4.abc.com

June 16:00:30 SNN CIB: [2219]: info:cib_ccm_msg_callback:PEER:snn.abc.com

June 16:00:31 SNN CRMD: [2223]: info:do_started:Delaying start, CCM (0000000000100000) not connected

June 16:00:31 SNN CRMD: [2223]: Info:mem_handle_event:Got an event oc_ev_ms_new_membership from CCM

June 16:00:31 SNN CRMD: [2223]: info:mem_handle_event:instance=5, nodes=2, new=2, lost=0, n_idx=0, new_idx=0, old_idx= 4

June 16:00:31 SNN CRMD: [2223]: Info:crmd_ccm_msg_callback:Quorum (re) attained after Event=new membership (id=5)

June 16:00:31 SNN CRMD: [2223]: Info:ccm_event_detail:NEW membership:trans=5, nodes=2, new=2, lost=0 n_idx=0, New_idx =0, old_idx=4

June 16:00:31 SNN CRMD: [2223]: info:ccm_event_detail: #011CURRENT: datanode4.abc.com [Nodeid=0, born=3]

June 16:00:31 SNN CRMD: [2223]: info:ccm_event_detail: #011CURRENT: snn.abc.com [nodeid=1, born=5]

June 16:00:31 SNN CRMD: [2223]: info:ccm_event_detail: #011NEW: datanode4.abc.com [Nodeid=0, born=3]

June 16:00:31 SNN CRMD: [2223]: info:ccm_event_detail: #011NEW: snn.abc.com [nodeid=1, born=5]

June 16:00:31 SNN CRMD: [2223]: Info:do_started:The local CRM is operational

June 16:00:31 SNN CRMD: [2223]: info:do_state_transition:State transition s_starting s_pending [input=i_pending Cause=c_ccm_callback origin=do_started]

4. View Cluster monitoring status

[Email protected] heartbeat]# Crm_mon

Refresh in 6s ...

============

Last Updated:fri June 19 16:11:34 2015

Current DC:snn.abc.com (6009ca6a-56eb-4d35-872e-3b8dc0fc9851)

2 Nodes configured.

0 Resources configured.

============

Node:datanode4.abc.com (0862D824-047E-4826-9E26-21A7603F53C8): Online

Node:snn.abc.com (6009ca6a-56eb-4d35-872e-3b8dc0fc9851): Online

4, the CRM command tool

[Email protected] heartbeat]# Crm_sh

/usr/sbin/crm_sh:31:deprecationwarning:the popen2 module is deprecated. Use the Subprocess module.

From Popen2 import Popen3

CRM # Help

USAGE:CRM (nodes|config|resources)

CRM # Nodes

CRM Nodes # Help

Usage:nodes (Status|list)

CRM Nodes # List

<node id= "0862d824-047e-4826-9e26-21a7603f53c8" uname= "datanode4.abc.com" type= "normal"/>

<node id= "6009ca6a-56eb-4d35-872e-3b8dc0fc9851" uname= "snn.abc.com" type= "normal"/>

CRM Nodes #

5, install heartbeat when automatically create a user hacluster, but no password, need to create

[Email protected] heartbeat]# cat/etc/passwd |grep hacluster

Hacluster:x:498:498:heartbeat User:/var/lib/heartbeat/cores/hacluster:/sbin/nologin

[Email protected] heartbeat]# passwd hacluster

Change the password for the user hacluster.

New Password:

Invalid Password: Too short

Invalid Password: too simple

Re-enter the new password:

PASSWD: All the authentication tokens have been successfully updated.

6, directly run Hb_gui

[Email protected] ~]# Hb_gui

Traceback (most recent):

File "/usr/bin/hb_gui", line $, in <module>

Import GTK, Gtk.glade, GObject

File "/usr/lib64/python2.6/site-packages/gtk-2.0/gtk/__init__.py", line @ <module>

_init ()

File "/usr/lib64/python2.6/site-packages/gtk-2.0/gtk/__init__.py", line-up, in _init

_gtk.init_check ()

Runtimeerror:could not open Display

There are error hints above

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/BB/wKiom1WD4kKjHhq7AAKYGhh-2CI498.jpg "title=" 22.png "alt=" Wkiom1wd4kkjhhq7aakyghh-2ci498.jpg "/>

Download and install Xmanager on the client

When you re-execute the command

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BB/wKiom1WD5IqhUc9SAATt458E-Gk900.jpg "title=" 23.png "alt=" Wkiom1wd5iqhuc9saatt458e-gk900.jpg "/>

Iii. definition of Ha_gui

1. Define the Master resource name

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BB/wKiom1WD6obRT9D7AAIGDmQ8Q0c090.jpg "title=" 24.png "alt=" Wkiom1wd6obrt9d7aaigdmq8q0c090.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/BB/wKiom1WD68-DYSRoAAMtZVcpp9A772.jpg "title=" 25.png "alt=" Wkiom1wd68-dysroaamtzvcpp9a772.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BB/wKiom1WD7P3Doq-HAAOUjNleYHs777.jpg "title=" 26.png "alt=" Wkiom1wd7p3doq-haaoujnleyhs777.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/B8/wKioL1WD7rvSGA57AAKp1DNjKL0275.jpg "title=" 27.png "alt=" Wkiol1wd7rvsga57aakp1dnjkl0275.jpg "/>

2, following the definition of master resources

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BC/wKiom1WD7zqSGPJRAAHxYAQLXVc864.jpg "title=" 28.png "alt=" Wkiom1wd7zqsgpjraahxyaqlxvc864.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/BC/wKiom1WD70ThbOgEAAMIoGhwXXQ196.jpg "title=" 29.png "alt=" Wkiom1wd70thbogeaamioghwxxq196.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/6E/B8/wKioL1WD8QTS3DV8AAKFjBQlP-o714.jpg "title=" 30.png "alt=" Wkiol1wd8qts3dv8aakfjbqlp-o714.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/B9/wKioL1WD9YXSnFgFAAJQK879AXg158.jpg "title=" 31.png "alt=" Wkiol1wd9yxsnfgfaajqk879axg158.jpg "/>

3, let two resources run the same node, there are two methods: (1) Define the arrangement constraints, (2) define the resource group

(1) Defining permutation constraints

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BC/wKiom1WD9heC77D6AAH_sfWkoJ0818.jpg "title=" 32.png "alt=" Wkiom1wd9hec77d6aah_sfwkoj0818.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/B9/wKioL1WD99eiBjFNAAOlnl00Zjw580.jpg "title=" 33.png "alt=" Wkiol1wd99eibjfnaaolnl00zjw580.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/6E/BC/wKiom1WD9oGgrfb2AAH4VGL2ac8178.jpg "title=" 34.png "alt=" Wkiom1wd9oggrfb2aah4vgl2ac8178.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/6E/BC/wKiom1WD9t_QLYzqAAEawH1atdk598.jpg "title=" 35.png "alt=" Wkiom1wd9t_qlyzqaaeawh1atdk598.jpg "/>

4, let the SNN node become prepared

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/BC/wKiom1WD98yS7sX-AAIv0RMduF4084.jpg "style=" float: none; "title=" 36.png "alt=" Wkiom1wd98ys7sx-aaiv0rmduf4084.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/6E/B9/wKioL1WD-X-BQLiUAAJdCkZBs-s135.jpg "style=" float: none; "title=" 37.png "alt=" Wkiol1wd-x-bqliuaajdckzbs-s135.jpg "/>

Iv. ways to define Groups

Web server:

vip:192.168.1.8

httpd

Nfs:/192.168.1.4:/web/htdocs hanging on to/var/www/html

1. Delete the original master resource

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/B9/wKioL1WD-_6BKl24AAHktqrP8ZI523.jpg "title=" 38.png "alt=" Wkiol1wd-_6bkl24aahktqrp8zi523.jpg "/>

2. Define Group master Source

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/BA/wKioL1WECZSBUX-BAAITS2IhDXw499.jpg "style=" float: none; "title=" 41.png "alt=" Wkiol1weczsbux-baaits2ihdxw499.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/BD/wKiom1WEB-KBPjLnAAMH5qhE7XI203.jpg "style=" float: none; "title=" 43.png "alt=" Wkiom1web-kbpjlnaamh5qhe7xi203.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/6E/BA/wKioL1WECZXiY09nAAI9tZyRdEE589.jpg "style=" float: none; "title=" 44.png "alt=" Wkiol1weczxiy09naai9tzyrdee589.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M00/6E/BD/wKiom1WEB-Oy5shJAAOr9bfv-Bg743.jpg "style=" float: none; "title=" 45.png "alt=" Wkiom1web-oy5shjaaor9bfv-bg743.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M02/6E/BD/wKiom1WEB-OCBJnKAAMwMXL56jM083.jpg "style=" float: none; "title=" 46.png "alt=" Wkiom1web-ocbjnkaamwmxl56jm083.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BA/wKioL1WECZbj9WRcAAJzBPG6uvc027.jpg "style=" float: none; "title=" 47.png "alt=" Wkiol1weczbj9wrcaajzbpg6uvc027.jpg "/>

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BA/wKioL1WECg3QYMOeAALWGkxiE-c668.jpg "title=" 48.png "alt=" Wkiol1wecg3qymoeaalwgkxie-c668.jpg "/>

3, httpd cannot start, check the log as follows

650) this.width=650; "src=" http://s3.51cto.com/wyfs02/M01/6E/BD/wKiom1WECJSzDz34AAtEM83nkEY497.jpg "title=" 49.png "alt=" Wkiom1wecjszdz34aatem83nkey497.jpg "/>

From the log, NFS normal hanging on to 4 this host, but httpd first start and close, I have not found out why, expect some altar know, may wish to point out, thank you all

This article is from the "Scattered People" blog, please be sure to keep this source http://zouqingyun.blog.51cto.com/782246/1663710

The heartbeat of High availability cluster resource management based on CRM (II.)