【正文】 pleteresourceinfo:heartbeat:19:47:23(none)DecacquisitionLocal[6852]:ndb217transitionremote[6852]:ndb217transitionremote[6852]:ndb217(pidgid498asStarting[6879]:ndb217/usr/lib/heartbeat/ipfailchildinfo:heartbeat:19:47:23Dec39。setstatusinfo:heartbeat:19:47:23activeDecstatusComm_now_up():[6852]:ndb217/etc/info:ndb217f~]2)此時(shí)在數(shù)據(jù)庫(kù)表中增加內(nèi)容之后啟動(dòng)剛剛關(guān)閉的ndbd，檢查新寫(xiě)入的數(shù)據(jù)是否已經(jīng)被同步過(guò)來(lái)。檢查ndbd關(guān)閉任意一臺(tái)ndbd的進(jìn)程，在ndb_mgm上查看是否失去連接。OK如果沒(méi)有出現(xiàn)異常，表明一切正常。IPaddr2[7291]:19:42:23Decstart/etc/invokedInvokingndb117RunningResourceManager[7196]:19:42:22DecldirectordExitingndb1175596withisforldirectord[7223]:19:42:22Decstatus/etc/invokedInvokingndb117LVSSyncDaemonSwap::masterndb1resourceinfo:ndb117OKiprequestrespiprequestresp[7175]:19:42:21iprequestrespDecRunningharc[7175]:19:42:21pleted.DecResourceinfo:heartbeat:19:42:21DecldirectordExitingndb1175596withisforldirectord[7153]:19:42:21Decstatus/etc/invokedInvokingndb117takeoverinfo:heartbeat:19:42:21ndb2.Decfortakeoverinfo:ndb117resourcesnice_failback:info:ndb117/etc/info:ndb117frombeinginfo:heartbeat:19:42:21.Dec39。noticeReceived[5462]:ndb117f]一切正常。anyconnectconnected,host)id=11fromaccepting(notanyconnectconnected,)id=9(Version:)id=7(Version:node(s)id=5)[mysqld(API)](Version:node(s)id=10)[ndb_mgmd(MGM)],0,2:1186ClusterServertondb_mgmManagementClusterndb_mgm所有檢查服務(wù)是否正常以下在ndb上執(zhí)行ndb_mgm[rootndb1101/77Mysqld164/26Ndbdcluster：順序：ndb_mgmdforever[rootsql2foreverscopeinet6lo:0scopebrdinetloscopeinet00:00:00:00:00:0000:00:00:00:00:00noqueue16436LOOPBACK,UP,LOWER_UPlo1:addrmysqlcluster]~]preferred_lftvalid_lfthost::1/128globalhostbrdlink/loopbackqdiscmtulo:shiplo查看lo:0[rootsql1/etc/sysconfig/networkscripts/ifcfglo\:0/etc/sysconfig/networkscripts/ifcfglo設(shè)置如何接收ARP包以下在所有RS上執(zhí)行~]prohrdhlendestinationipsourceipinbytesbytes)packets,ACCEPTFORWARD0000/00000000/00000000/0000any00/0000/00eth0*0prohrdhlendestinationipsourceipinbytesbytes)packets,ACCEPTOUT0000/00000000/00000000/0000any00/0000/00DROPopdestinationhwsourcehwouttargetpkts823K29375(policynChainL~]mysqlcluster]prohrdhlendestinationipsourceipinbytesbytes)packets,ACCEPTFORWARD0000/00000000/00000000/0000any00/0000/00eth0*0prohrdhlendestinationipsourceipinbytesbytes)packets,ACCEPTOUT0000/00000000/00000000/0000any00/0000/00DROPopdestinationhwsourcehwouttargetpkts819K29243(policynChainLmysqlcluster]on/etc/2345save/sbin/chkconfig/etc/以下在RS服務(wù)器上執(zhí)行：ARP轉(zhuǎn)發(fā)限制MD或者BD采用ARP欺騙將ARP包轉(zhuǎn)發(fā)給下面的realserver。~]pid:statusmaster/etc/][rootndb2pid:statusmaster/etc/在MD和BD上檢查L(zhǎng)VSSyncDaemonSwap的狀態(tài)：[rootndb1InActConn[rootndb2WeightRemoteAddress:PortScheduler(size=4096)ProtversionVirtualL~]]Route303431wrrActiveConnForwardFlagsLocalAddress:PortServernIP/sbin/ipvsadm利用ipvs檢查包轉(zhuǎn)發(fā)是否生效[rootndb1/etc/[rootndb2stoppedstatusldirectord/usr/sbin/ldirectord[rootndb25596[rootndb1withisfor]~]現(xiàn)在在MD（164）上已經(jīng)生效了。preferred_lftvalid_lftlinkfe80::230:48ff:fe28:c4af/64globalbrdlink/ether1000pfifo_fast1500BROADCAST,MULTICAST,UP,LOWER_UPeth02:addr~]]preferred_lftvalid_lftlinkfe80::230:48ff:fe28:c685/64secondaryscopebrdineteth0scopebrdinetff:ff:ff:ff:ff:ff00:30:48:28:c6:85qlenqdiscmtueth0:shipeth0[rootndb1addrstop/etc/delheartbeatlevel600:wq設(shè)置屬性并使heartbeat開(kāi)機(jī)啟動(dòng)LVSSyncDaemonSwap::master\vimd5authvigid=haclient/usr/lib/heartbeat/ipfailapiauthndb2respawnndb1nodeoffnode1eth1mcastlocal0bcastlogfacility配置heartbeatHeartbeat有3個(gè)配置文件：AuthkeysHaresourcesldirectord進(jìn)程的配置文件一共需要配置4個(gè)配置文件。心跳線使用eth1口，用交叉線連接起來(lái)。shell這樣安裝的perl包不知道為何不好使？奇怪這里VIP實(shí)際上是綁定在2臺(tái)director上。CPANinstallUvhheartbeatRpmUvh1使修改生效：/sbin/sysctl改為：=/etc/modulesip_vs_dhip_vs_ftpip_vsip_vs_lblcip_vs_lblcrip_vs_lcip_vs_nqip_vs_rrip_vs_sedip_vs_ship_vs_wlcip_vs_wrr:wqVi:wqchmodip_vs_wlcmodprobeip_vs_sedmodprobeip_vs_nqmodprobeip_vs_lblcrmodprobeip_vsmodprobeip_vs_dhmodprobe在MD和BD修改IP轉(zhuǎn)發(fā)：vi等等以下操作在所有服務(wù)器上執(zhí)行：主機(jī)名確認(rèn)：分別執(zhí)行：unameserverBD3)BackupDirector（分發(fā)器）Heartbeat中包含以下幾部分：1)[rootndb1{printawk*.rpmls我下載了如下的軟件包：[rootndb1heartbeatCdultramonkey（）首先在NDB1（164）和NDB2（26）上下載heartbeat的軟件包：下載所有的rpm包：Cd所以，我們現(xiàn)在在2臺(tái)API之間做LVS。即便做了API的判斷和選擇，因?yàn)闆](méi)有實(shí)現(xiàn)負(fù)載均衡，服務(wù)器的性能沒(méi)有充分利用。cluster做好以后，數(shù)據(jù)庫(kù)分別建立同名的數(shù)據(jù)庫(kù)以后，權(quán)限分配好，然后只要在一臺(tái)上寫(xiě)入數(shù)據(jù)，其他的NDB就存儲(chǔ)了相同的數(shù)據(jù)。cluster已經(jīng)可以正常工作了。test。*MysqlusepxxxxxxxxxxxxMysql以上做完以后，可以通過(guò)任意一臺(tái)API上創(chuàng)建表，并寫(xiě)數(shù)據(jù)到表中，其他數(shù)據(jù)庫(kù)都會(huì)同步寫(xiě)入。MysqlintoMysqltest(intcreateprivileges。Mysql39。identified.%39。toongranttestdatabase。createpxxxxxxxxxxxxMysql的API，所以創(chuàng)建數(shù)據(jù)庫(kù)的時(shí)候，都需要?jiǎng)?chuàng)建一遍。du/var/lib/mysqlcluster/ndb_4_fs/[rootsql2~]ndbd由于77只有2GB的內(nèi)存，所以，加上mysqld用掉的，2GB的內(nèi)存似乎已經(jīng)所剩無(wú)幾了。R15rootmysqld223952119m0mysql%CPUSHRRESVIRTNIPRUSER1400812kfree,0ktotal,68256kfree,2005868ktotal,%stMem:%si,%hi,%wa,%id,0078running,total,average:11:59,up~]可能啟動(dòng)的時(shí)候會(huì)有警告，但是沒(méi)關(guān)系的。按照1：，需要4GB的內(nèi)存。配置完成。onOK，到此mysql2345/etc/Chkconfig/etc/Chmod39。755/var/lib/mysqlcluster/39。ndb_mgmd將服務(wù)增加到開(kāi)機(jī)啟動(dòng)服務(wù)項(xiàng)中：164/26:echohost)ndb_mgmfromaccepting(notanyconnectconnected,host)id=10fromaccepting(not(Version:)id=8(Version:)id=67(Version:)id=22Node