Centos7上利用corosync+pacemaker+crmsh構建高可用叢集
阿新 • • 發佈：2018-11-16
一、高可用叢集框架
資源型別：
primitive(native)：表示主資源 group：表示組資源，組資源裡包含多個主資源 clone：表示克隆資源 masterslave：表示主從資源
資源約束方式：
位置約束：定義資源對節點的傾向性 排序約束：定義資源彼此能否執行在同一節點的傾向性 順序約束：多個資源啟動順序的依賴關係
HA叢集常用的工作模型：
AP：兩節點，activepassive，工作於主備模型 AA：兩節點，activeactive，工作於主主模型 N-M：NM，N個節點，M個服務，假設每個節點執行一個服務，活動節點數為N，備用節點數為N-M
在叢集分裂(split-brain)時需要使用到資源隔離，有兩種隔離級別：
STONITH：節點級別的隔離，通過斷開一個節點的電源或者重新啟動節點 fencing：資源級別的隔離，類似通過向交換機發出隔離訊號，特意讓資料無法通過此介面 當叢集分裂，即分裂後的一個叢集的法定票數小於總票數一半時採取對資源的控制策略，
二、在centos7上建立Ha cluster
centos7(corosync v2 + pacemaker)
叢集的全生命週期管理工具：
pcs agent(pcsd) crmsh agentless (pssh)
1、叢集配置前提
時間同步，基於當前正在使用的主機名互相訪問，是否會用到仲裁裝置
#更改主機名，兩臺主機的都要修改（192.168.1.114為ns2.xinfeng.com）（192.168.1.113為ns3.xinfeng.com） [[email protected] ~]# hostnamectl set-hostname ns2.xinfeng.com [[email protected] ~]# uname -n ns2.xinfeng.com [[email protected] ~]# vim etchosts 192.168.1.114 ns2.xinfeng.com 192.168.1.113 ns3.xinfeng.com #同步時間 [[email protected] ~]# ntpdate s1a.time.edu.cn [[email protected] ~]# ssh 192.168.1.113 'date';date The authenticity of host '192.168.1.113 (192.168.1.113)' can't be established. ECDSA key fingerprint is 09f9398c354dba2d134f3c9cb15854ec. Are you sure you want to continue connecting (yesno) yes Warning Permanently added '192.168.1.113' (ECDSA) to the list of known hosts. [email protected]'s password 2016年 05月 28日 星期六 131807 CST 2016年 05月 28日 星期六 131807 CST
2、安裝pcs並啟動叢集
#192.168.1.113 [[email protected] ~]# yum install pcs #192.168.1.114 [[email protected] ~]# yum install pcs #用ansible開起服務並使服務開機執行 [[email protected] ~]# vim etcansiblehosts [ha] 192.168.1.114 192.168.1.113 [[email protected] ~]# ansible ha -m service -a 'name=pcsd state=started enabled=yes' 192.168.1.113 SUCCESS = { changed false, enabled true, name pcsd, state started } 192.168.1.114 SUCCESS = { changed true, enabled true, name pcsd, state started } #用ansible檢視服務是否啟動 [[email protected] ~]# ansible ha -m shell -a 'systemctl status pcsd' 192.168.1.114 SUCCESS rc=0 ● pcsd.service - PCS GUI and remote configuration interface Loaded loaded (usrlibsystemdsystempcsd.service; enabled; vendor preset disabled) Active active (running) since 六 2016-05-28 133619 CST; 2min 32s ago Main PID 2736 (pcsd) CGroup system.slicepcsd.service ├─2736 binsh usrlibpcsdpcsd start ├─2740 binbash -c ulimit -S -c 0 devnull 2&1 ; usrbinruby -Iusrlibpcsd usrlibpcsdssl.rb └─2741 usrbinruby -Iusrlibpcsd usrlibpcsdssl.rb 5月 28 133616 ns2.xinfeng.com systemd[1] Starting PCS GUI and remote configuration interface... 5月 28 133619 ns2.xinfeng.com systemd[1] Started PCS GUI and remote configuration interface. 192.168.1.113 SUCCESS rc=0 ● pcsd.service - PCS GUI and remote configuration interface Loaded loaded (usrlibsystemdsystempcsd.service; enabled; vendor preset disabled) Active active (running) since 六 2016-05-28 133526 CST; 3min 24s ago Main PID 2620 (pcsd) CGroup system.slicepcsd.service ├─2620 binsh usrlibpcsdpcsd start ├─2624 binbash -c ulimit -S -c 0 devnull 2&1 ; usrbinruby -Iusrlibpcsd usrlibpcsdssl.rb └─2625 usrbinruby -Iusrlibpcsd usrlibpcsdssl.rb 5月 28 133524 ns3.xinfeng.com systemd[1] Starting PCS GUI and remote configuration interface... 5月 28 133526 ns3.xinfeng.com systemd[1] Started PCS GUI and remote configuration interface. #給hacluster使用者增加密碼 [[email protected] ~]# ansible ha -m shell -a 'echo 123 passwd --stdin hacluster' 192.168.1.113 SUCCESS rc=0 更改使用者 hacluster 的密碼 。 passwd：所有的身份驗證令牌已經成功更新。 192.168.1.114 SUCCESS rc=0 更改使用者 hacluster 的密碼 。 passwd：所有的身份驗證令牌已經成功更新。 #認證節點身份，使用者名稱和密碼為上面設定的hacluster和123，注意iptables規則，否則會出現無法聯絡的情況 [[email protected] ~]# pcs cluster auth ns2.xinfeng.com ns3.xinfeng.com Username hacluster Password ns2.xinfeng.com Authorized ns3.xinfeng.com Authorized #配置叢集，叢集名字為xinfengcluster，叢集中有2個節點 [[email protected] ~]# pcs cluster setup --name xinfengcluster ns2.xinfeng.com ns3.xinfeng.com Shutting down pacemakercorosync services... Redirecting to binsystemctl stop pacemaker.service Redirecting to binsystemctl stop corosync.service Killing any remaining services... Removing all cluster configuration files... ns2.xinfeng.com Succeeded ns3.xinfeng.com Succeeded Synchronizing pcsd certificates on nodes ns2.xinfeng.com, ns3.xinfeng.com... ns2.xinfeng.com Success ns3.xinfeng.com Success Restaring pcsd on the nodes in order to reload the certificates... ns2.xinfeng.com Success ns3.xinfeng.com Success #檢視下配置檔案 [[email protected] ~]# cat etccorosynccorosync.conf totem { #叢集的資訊 version 2 #版本 secauth off #安全功能是否開起 cluster_name xinfengcluster #叢集名稱 transport udpu #傳輸協議udpu也可以設定為udp } nodelist { #叢集中的所有節點 node { ring0_addr ns2.xinfeng.com nodeid 1 #節點ID } node { ring0_addr ns3.xinfeng.com nodeid 2 } } quorum { #仲裁投票 provider corosync_votequorum #投票系統 two_node 1 #是否為2節點叢集 } logging { #日誌 to_logfile yes #是否記錄日誌 logfile varlogclustercorosync.log #日誌檔案位置 to_syslog yes #是否記錄系統日誌 } #啟動叢集 [[email protected] ~]# pcs cluster start --all ns3.xinfeng.com Starting Cluster... ns2.xinfeng.com Starting Cluster... #檢視ns2.xinfeng.com節點是否啟動 [[email protected] ~]# corosync-cfgtool -s Printing ring status. Local node ID 1 RING ID 0 id = 192.168.1.114 status = ring 0 active with no faults #檢視ns3.xinfeng.com節點是否啟動 [[email protected] ~]# corosync-cfgtool -s Printing ring status. Local node ID 2 RING ID 0 id = 192.168.1.113 status = ring 0 active with no faults #檢視叢集資訊 [[email protected] ~]# corosync-cmapctl grep members runtime.totem.pg.mrp.srp.members.1.config_version (u64) = 0 runtime.totem.pg.mrp.srp.members.1.ip (str) = r(0) ip(192.168.1.114) runtime.totem.pg.mrp.srp.members.1.join_count (u32) = 1 runtime.totem.pg.mrp.srp.members.1.status (str) = joined runtime.totem.pg.mrp.srp.members.2.config_version (u64) = 0 runtime.totem.pg.mrp.srp.members.2.ip (str) = r(0) ip(192.168.1.113) runtime.totem.pg.mrp.srp.members.2.join_count (u32) = 1 runtime.totem.pg.mrp.srp.members.2.status (str) = joined [[email protected] ~]# pcs status Cluster name xinfengcluster WARNING no stonith devices and stonith-enabled is not false Last updated Sat May 28 143823 2016 Last change Sat May 28 143315 2016 by hacluster via crmd on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum #DC為全域性仲裁節點 2 nodes and 0 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources PCSD Status ns2.xinfeng.com Online ns3.xinfeng.com Online Daemon Status corosync activedisabled pacemaker activedisabled pcsd activeenabled
3、使用crmsh配置叢集
安裝opensuse上的yum源
centos6
cd etcyum.repos.d wget httpdownload.opensuse.orgrepositoriesnetworkha-clusteringStableCentOS_CentOS-6networkha-clusteringStable.repo cd yum -y install crmsh
centos7
cd etcyum.repos.d wget httpdownload.opensuse.orgrepositoriesnetworkha-clusteringStableCentOS_CentOS-7networkha-clusteringStable.repo cd yum -y install crmsh
在其中一臺主機上配置crmsh（192.1681.1.114）
#顯示當前的叢集狀態 [[email protected] yum.repos.d]# crm status Last updated Sat May 28 165252 2016 Last change Sat May 28 143315 2016 by hacluster via crmd on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 0 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources
4、兩個節點上分別裝上httpd
[[email protected] ~]# ansible ha -m shell -a 'yum install httpd -y' [[email protected] ~]# echo h1ns2.xinfeng.comh1 varwwwhtmlindex.html [[email protected] ~]# echo h1ns3.xinfeng.comh1 varwwwhtmlindex.html #測試能否正常啟動，頁面是否正常 [[email protected] ~]# ansible ha -m service -a 'name=httpd state=started enabled=yes' #centos6必須關閉服務，關閉開機啟動，之後服務會變成資源，所以請確保服務不啟動也不開機啟動 [[email protected] ~]# ansible ha -m service -a 'name=httpd state=stopped enabled=no' #centos7必須關閉服務，開起開機啟動 [[email protected] ~]# ansible ha -m service -a 'name=httpd state=stopped enabled=yes'
5、配置叢集
VIP為192.168.1.91，服務是httpd，將VIP和httpd作為資源來進行配置
[[email protected] ~]# crm crm(live)# ra #進入資源代理 crm(live)ra# classes #檢視可以代理的資源型別 lsb ocf .isolation heartbeat openstack pacemaker service stonith systemd crm(live)ra# list systemd #檢視systemd型別可代理的服務，其中有httpd NetworkManager NetworkManager-wait-online auditd brandbot corosync cpupower crond dbus display-manager dm-event dracut-shutdown ebtables emergency exim firewalld [email protected] httpd ip6tables iptables irqbalance kdump kmod-static-nodes ldconfig libvirtd lvm2-activation lvm2-lvmetad lvm2-lvmpolld lvm2-monitor [email protected] microcode network pacemaker pcsd plymouth-quit plymouth-quit-wait plymouth-read-write plymouth-start polkit postfix rc-local rescue rhel-autorelabel rhel-autorelabel-mark rhel-configure rhel-dmesg rhel-import-state rhel-loadmodules rhel-readonly rsyslog sendmail sshd sshd-keygen syslog systemd-ask-password-console systemd-ask-password-plymouth systemd-ask-password-wall systemd-binfmt systemd-firstboot systemd-fsck-root systemd-hwdb-update systemd-initctl systemd-journal-catalog-update systemd-journal-flush systemd-journald systemd-logind systemd-machine-id-commit systemd-modules-load systemd-random-seed systemd-random-seed-load systemd-readahead-collect systemd-readahead-done systemd-readahead-replay systemd-reboot systemd-remount-fs systemd-shutdownd systemd-sysctl systemd-sysusers systemd-tmpfiles-clean systemd-tmpfiles-setup systemd-tmpfiles-setup-dev systemd-udev-trigger systemd-udevd systemd-update-done systemd-update-utmp systemd-update-utmp-runlevel systemd-user-sessions systemd-vconsole-setup tuned wpa_supplicant crm(live)ra# cd crm(live)# configure #配置資源，資源名為webip，ip為192.168.1.91 crm(live)configure# primitive webip ocfheartbeatIPaddr params ip=192.168.1.91 crm(live)configure# show node 1 ns2.xinfeng.com node 2 ns3.xinfeng.com primitive webip IPaddr params ip=192.168.1.91 property cib-bootstrap-options have-watchdog=false dc-version=1.1.13-10.el7_2.2-44eb2dd cluster-infrastructure=corosync cluster-name=xinfengcluster crm(live)configure# verify #校驗，因為沒有隔離裝置所以報錯 ERROR error unpack_resources Resource start-up disabled since no STONITH resources have been defined error unpack_resources Either configure some or disable STONITH with the stonith-enabled option error unpack_resources NOTE Clusters with shared data need STONITH to ensure data integrity Errors found during check config not valid crm(live)configure# property stonith-enabled=false #關閉隔離裝置的設定 crm(live)configure# verify #再次校驗 crm(live)configure# commit #提交，是配置生效 crm(live)configure# cd crm(live)# status Last updated Sat May 28 174146 2016 Last change Sat May 28 174131 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 1 resource configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns2.xinfeng.com #VIP已經啟動在了ns2上 crm(live)# quit bye [[email protected] ~]# ip addr #驗證一下VIP是否已經配置到網絡卡上 1 lo LOOPBACK,UP,LOWER_UP mtu 65536 qdisc noqueue state UNKNOWN linkloopback 000000000000 brd 000000000000 inet 127.0.0.18 scope host lo valid_lft forever preferred_lft forever inet6 1128 scope host valid_lft forever preferred_lft forever 2 eno16777728 BROADCAST,MULTICAST,UP,LOWER_UP mtu 1500 qdisc pfifo_fast state UP qlen 1000 linkether 000c299157d1 brd ffffffffffff inet 192.168.1.11424 brd 192.168.1.255 scope global dynamic eno16777728 valid_lft 5918sec preferred_lft 5918sec inet 192.168.1.9124 brd 192.168.1.255 scope global secondary eno16777728 valid_lft forever preferred_lft forever inet6 fe8020c29fffe9157d164 scope link valid_lft forever preferred_lft forever #將當前節點切換為備用節點 [[email protected] ~]# crm crm(live)# node crm(live)node# standby crm(live)node# cd crm(live)# status Last updated Sat May 28 174541 2016 Last change Sat May 28 174534 2016 by root via crm_attribute on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 1 resource configured Node ns2.xinfeng.com standby Online [ ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns3.xinfeng.com #資源重新上線 crm(live)# node online crm(live)# status Last updated Sat May 28 174640 2016 Last change Sat May 28 174637 2016 by root via crm_attribute on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 1 resource configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns3.xinfeng.com #配置httpd資源，資源名為httpd [[email protected] ~]# crm crm(live)# configure crm(live)configure# primitive webser systemdhttpd crm(live)configure# verify crm(live)configure# commit crm(live)configure# cd crm(live)# status Last updated Sat May 28 175015 2016 Last change Sat May 28 174956 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 2 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns3.xinfeng.com webser (systemdhttpd) Started ns2.xinfeng.com #將兩個資源放在webhttp組中，資源啟動順序是webip，webser crm(live)# configure crm(live)configure# group webhttp webip webser crm(live)configure# verify crm(live)configure# commit crm(live)configure# cd crm(live)# status Last updated Sat May 28 175248 2016 Last change Sat May 28 175241 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 2 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources Resource Group webhttp webip (ocfheartbeatIPaddr) Started ns3.xinfeng.com webser (systemdhttpd) Started ns3.xinfeng.com
由於是2個節點，會存在法定票數不足導致的資源不轉移的情況，解決此問題的方法有四種：
1、可以增加一個ping node節點。 2、可以增加一個仲裁磁碟。 3、讓叢集中的節點數成奇數個。 4、直接忽略當叢集沒有法定票數時直接忽略。
這裡我用的是第四種方式
[[email protected] ~]# crm crm(live)# configure crm(live)configure# property no-quorum-policy=ignore crm(live)configure# verify crm(live)configure# commit
這樣做還不夠，因為沒有對資源進行監控，所以資源出現問題依然不會轉移
現在只能測試下資源是否在ns3.xinfeng.com上啟動了
要對資源進行監控需要在全域性下命令primitive定義資源時一同定義，因此先把之前定義的資源刪掉後重新定義
[[email protected] ~]# crm crm(live)# resource crm(live)resource# show Resource Group webhttp webip (ocfheartbeatIPaddr) Started webser (systemdhttpd) Started crm(live)resource# stop webhttp #停掉所有資源 crm(live)resource# show Resource Group webhttp webip (ocfheartbeatIPaddr) (target-roleStopped) Stopped webser (systemdhttpd) (target-roleStopped) Stopped crm(live)configure# edit #編輯資源定義配置檔案 node 1 ns2.xinfeng.com attributes standby=off node 2 ns3.xinfeng.com primitive webip IPaddr #刪除 params ip=192.168.1.91 #刪除 primitive webser systemdhttpd #刪除 group webhttp webip webser #刪除 meta target-role=Stopped #刪除 property cib-bootstrap-options have-watchdog=false dc-version=1.1.13-10.el7_2.2-44eb2dd cluster-infrastructure=corosync cluster-name=xinfengcluster stonith-enabled=false no-quorum-policy=ignore crm(live)configure# verify crm(live)configure# commit crm(live)configure# cd crm(live)# status Last updated Sat May 28 233201 2016 Last change Sat May 28 233149 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 0 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources
6、重新定義帶有監控的資源
crm(live)# configure #每60秒監控一次，超時時長為20秒，時間不能小於建議時長，否則會報錯 crm(live)configure# primitive webip ocfIPaddr params ip=192.168.1.91 op monitor timeout=20s interval=60s crm(live)configure# primitive webser systemdhttpd op monitor timeout=20s interval=60s crm(live)configure# group webhttp webip webser crm(live)configure# property no-quorum-policy=ignore crm(live)configure# verify crm(live)configure# commit crm(live)configure# cd crm(live)# status Last updated Sat May 28 234103 2016 Last change Sat May 28 234036 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns2.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 2 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources Resource Group webhttp webip (ocfheartbeatIPaddr) Started ns2.xinfeng.com webser (systemdhttpd) Started ns2.xinfeng.com
測試一下，將服務停掉，20秒後服務又自動會啟動
[[email protected] ~]# ip addr 1 lo LOOPBACK,UP,LOWER_UP mtu 65536 qdisc noqueue state UNKNOWN linkloopback 000000000000 brd 000000000000 inet 127.0.0.18 scope host lo valid_lft forever preferred_lft forever inet6 1128 scope host valid_lft forever preferred_lft forever 2 eno16777728 BROADCAST,MULTICAST,UP,LOWER_UP mtu 1500 qdisc pfifo_fast state UP qlen 1000 linkether 000c299157d1 brd ffffffffffff inet 192.168.1.11424 brd 192.168.1.255 scope global dynamic eno16777728 valid_lft 6374sec preferred_lft 6374sec inet 192.168.1.9124 brd 192.168.1.255 scope global secondary eno16777728 valid_lft forever preferred_lft forever inet6 fe8020c29fffe9157d164 scope link valid_lft forever preferred_lft forever [[email protected] ~]# service httpd stop Redirecting to binsystemctl stop httpd.service
編輯配置檔案，隨便亂加幾行讓服務不能啟動
[[email protected] ~]# vim etchttpdconfhttpd.conf [[email protected] ~]# service httpd stop Redirecting to binsystemctl stop httpd.service
服務成功切換到ns3上
7、【注意】當重新恢復httpd服務後記得清除資源的錯誤資訊，否則無法啟動資源
crm(live)# resource crm(live)resource# cleanup webser #清楚webser之前的錯誤資訊 Cleaning up webser on ns2.xinfeng.com, removing fail-count-webser Cleaning up webser on ns3.xinfeng.com, removing fail-count-webser Waiting for 2 replies from the CRMd.. OK crm(live)resource# show webip (ocfheartbeatIPaddr) Started webser (systemdhttpd) Started
8、定義資源約束
刪除組資源
[[email protected] ~]# crm crm(live)# configure crm(live)configure# delete webhttp #刪除webhttp組 crm(live)configure# verify crm(live)configure# commit crm(live)configure# cd crm(live)# status Last updated Sun May 29 092257 2016 Last change Sun May 29 092248 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns3.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 2 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns2.xinfeng.com webser (systemdhttpd) Started ns3.xinfeng.com
排列約束
[[email protected] ~]# crm crm(live)# configure crm(live)configure# colocation webser_with_webip inf webser webip #定義webser和webip兩個資源必須在一起 crm(live)configure# verify crm(live)configure# commit crm(live)configure# cd crm(live)# status Last updated Sun May 29 094036 2016 Last change Sun May 29 094028 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns3.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 2 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns2.xinfeng.com #可以看到兩個資源都在ns2上啟動了 webser (systemdhttpd) Started ns2.xinfeng.com
順序約束
crm(live)# configure #webip先於webser啟動，強制的先啟動webip在啟動webser crm(live)configure# order webip_before_webser mandatory webip webser crm(live)configure# verify crm(live)configure# commit crm(live)configure# show xml #檢視之前詳細定義
位置約束
#先看下當前的位置 crm(live)# status Last updated Sun May 29 094842 2016 Last change Sun May 29 094435 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns3.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 2 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns2.xinfeng.com webser (systemdhttpd) Started ns2.xinfeng.com #定義位置約束讓資源更傾向於ns3上 crm(live)# configure 定義一個webservice的位置約束在節點2上，資源webip對ns3的傾向性是100 crm(live)configure# location webservice_pref_node2 webip 100 ns3.xinfeng.com crm(live)configure# verify crm(live)configure# commit crm(live)configure# cd crm(live)# status Last updated Sun May 29 101212 2016 Last change Sun May 29 101204 2016 by root via cibadmin on ns2.xinfeng.com Stack corosync Current DC ns3.xinfeng.com (version 1.1.13-10.el7_2.2-44eb2dd) - partition with quorum 2 nodes and 2 resources configured Online [ ns2.xinfeng.com ns3.xinfeng.com ] Full list of resources webip (ocfheartbeatIPaddr) Started ns3.xinfeng.com #很明顯，資源都轉移到了ns3上了 webser (systemdhttpd) Started ns3.xinfeng.com
三、總結
1、當重新恢復資源的服務後一定記得清除資源的錯誤資訊，否則無法啟動資源
2、在利用corosync+pacemaker且是兩個節點實現高可用時，需要注意的是要設定全域性屬性把stonith裝置關閉，忽略法定票數不大於一半的機制
3、注意selinux和iptables對服務的影響
4、注意節點相互用etchosts來解析
5、節點時間一定要保持同步
6、節點相互間進行無金鑰通訊