Oracle RAC故障分析与处理
更多精彩内容尽在www.leonarding.com《Oracle RAC故障分析与处理》
一RAC环境
RAC架构，2节点信息
节点1
SQL> show parameter instance
NAME TYPE VALUE
------------------------------------ ----------- -----------------------------------------------
active_instance_count integer
cluster_database_instances integer 2
instance_groups string
instance_name string RACDB1
instance_number Integer 1
instance_type string RDBMS
open_links_per_instance integer 4
parallel_instance_group string
parallel_server_instances integer 2
节点2
SQL> show parameter instance
NAME TYPE VALUE
------------------------------------ ----------- ------------------------------------------
active_instance_count integer
cluster_database_instances integer 2
instance_groups string
instance_name string RACDB2
instance_number integer 2
instance_type string RDBMS
open_links_per_instance integer 4
parallel_instance_group string
parallel_server_instances integer 2
数据库版本
SQL> select * from v$version;
BANNER
----------------------------------------------------------------
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Prod
PL/SQL Release 10.2.0.1.0 - Production
CORE 10.2.0.1.0 Production
TNS for Linux: Version 10.2.0.1.0 - Production
NLSRTL Version 10.2.0.1.0 - Production
操作系统信息
节点1
$ uname -a
Linux rac1 2.6.18-53.el5 #1 SMP Wed Oct 10 16:34:02 EDT 2007 i686 i686 i386 GNU/Linux
节点2
$ uname -a
Linux rac2 2.6.18-53.el5 #1 SMP Wed Oct 10 16:34:02 EDT 2007 i686 i686 i386 GNU/Linux
RAC所有资源信息
$ crs_stat -t
Name Type Target State Host
----------------------------------------------------------------------------------------------
ora....B1.inst application ONLINE ONLINE rac1
ora....B2.inst application ONLINE ONLINE rac2
ora....DB1.srv application ONLINE ONLINE rac2
ora.....TAF.cs application ONLINE ONLINE rac2
ora.RACDB.dbapplication ONLINE ONLINE rac2
ora....SM1.asmapplication ONLINE ONLINE rac1
ora....C1.lsnr application ONLINE ONLINE rac1
ora.rac1.gsd application ONLINE ONLINE rac1
ora.rac1.ons application ONLINE ONLINE rac1
ora.rac1.vip application ONLINE ONLINE rac1
ora....SM2.asmapplication ONLINE ONLINE rac2
ora....C2.lsnr application ONLINE ONLINE rac2
ora.rac2.gsd application ONLINE ONLINE rac2
ora.rac2.ons application ONLINE ONLINE rac2
ora.rac2.vip application ONLINE ONLINE rac2
二模拟两个节点内联网不通，观察RAC会出现什么现象？给出故障定位的整个过程
本小题会模拟RAC的私有网络不通现象，然后定位故障原因，最后排除故障。
1.首先RAC是一个非常健康的状态
$ crs_stat -t
Name Type Target State Host
----------------------------------------------------------------------------------------------
ora....B1.inst application ONLINE ONLINE rac1
ora....B2.inst application ONLINE ONLINE rac2
ora....DB1.srv application ONLINE ONLINE rac2
ora.....TAF.cs application ONLINE ONLINE rac2
ora.RACDB.dbapplication ONLINE ONLINE rac2
ora....SM1.asmapplication ONLINE ONLINE rac1
ora....C1.lsnr application ONLINE ONLINE rac1
ora.rac1.gsd application ONLINE ONLINE rac1
ora.rac1.ons application ONLINE ONLINE rac1
ora.rac1.vip application ONLINE ONLINE rac1
ora....SM2.asmapplication ONLINE ONLINE rac2
ora....C2.lsnr application ONLINE ONLINE rac2
ora.rac2.gsd application ONLINE ONLINE rac2
ora.rac2.ons application ONLINE ONLINE rac2
ora.rac2.vip application ONLINE ONLINE rac2
检查CRS进程状态（CRSCSSEVM）
$ crsctl check crs
CSS appears healthy
CRS appears healthy
EVM appears healthy
检查OCR磁盘状态，没有问题
$ ocrcheck
Status of Oracle Cluster Registry is as follows :
Version : 2
Total space (kbytes) : 104344
Used space (kbytes) : 4344
Available space (kbytes) : 100000
ID : 1752469369
Device/File Name : /dev/raw/raw1
Device/File integrity check succeeded
Device/File not configured
Cluster registry integrity check succeeded
检查vote disk状态
$ crsctl query css votedisk
0. 0 /dev/raw/raw2 显示2号裸设备为表决磁盘
located 1 votedisk(s). 只定位1个表决磁盘
2.手工禁用一个私有网卡
$ cat /etc/hosts
127.0.0.1 localhost.localdomain localhost
::1 localhost6.localdomain6 localhost6
##Public Network - (eth0)
##Private Interconnect - (eth1)
##Public Virtual IP (VIP) addresses - (eth0)
192.168.1.101 rac1 这是RAC的共有网卡
192.168.1.102 rac2
192.168.2.101 rac1-priv 这是RAC的私有网卡
192.168.2.102 rac2-priv
192.168.1.201 rac1-vip 这是RAC虚拟网卡
192.168.1.202 rac2-vip
看一下IP地址和网卡的对应关系
$ ifconfig
eth0 Link encap:EthernetHWaddr 00:0C:29:8F:F1:87
inet addr:192.168.1.102Bcast:192.168.1.255Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe8f:f187/64 Scope:Link
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
RX packets:360 errors:0 dropped:0 overruns:0 frame:0
TX packets:593 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:46046 (44.9 KiB)TX bytes:62812 (61.3 KiB)
Interrupt:185 Base address:0x14a4
eth0:1 Link encap:EthernetHWaddr 00:0C:29:8F:F1:87
inet addr:192.168.1.202Bcast:192.168.1.255Mask:255.255.255.0
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
Interrupt:185 Base address:0x14a4
eth1 Link encap:EthernetHWaddr 00:0C:29:8F:F1:91
inet addr:192.168.2.102Bcast:192.168.2.255Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe8f:f191/64 Scope:Link
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
RX packets:76588 errors:0 dropped:0 overruns:0 frame:0
TX packets:58002 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:65185420 (62.1 MiB)TX bytes:37988820 (36.2 MiB)
Interrupt:193 Base address:0x1824
eth2 Link encap:EthernetHWaddr 00:0C:29:8F:F1:9B
inet addr:192.168.203.129Bcast:192.168.203.255Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe8f:f19b/64 Scope:Link
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
RX packets:339 errors:0 dropped:0 overruns:0 frame:0
TX packets:83 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:42206 (41.2 KiB)TX bytes:10199 (9.9 KiB)
Interrupt:169 Base address:0x18a4
lo Link encap:Local Loopback
inet addr:127.0.0.1Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNINGMTU:16436Metric:1
RX packets:99403 errors:0 dropped:0 overruns:0 frame:0
TX packets:99403 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:18134658 (17.2 MiB)TX bytes:18134658 (17.2 MiB)
eth 0 对应RAC的共有网卡
eth 1 对应RAC的私有网卡
eth0:1对应RAC的虚拟网卡
我们现在禁止eth1私有网卡来完成内联网网络不通现象，方法很简单
ifdown eth1 禁用网卡
ifup eth1 激活网卡
$ su – root 需要使用root用户哦，否则提示Users cannot control this device.
Password:
# ifdown eth1
我从17:18:51敲入这个命令，4分钟之后节点2重启，大家知道发生了什么现象嘛？
Good 这就是传说中RAC脑裂brain split问题，当节点间的内联网不通时，无法信息共享，就会出现脑裂现象，RAC必须驱逐其中一部分节点来保护数据的一致性，被驱逐的节点被强制重启，这不节点2自动重启了么。又说回来，那为什么节点2重启，其他节点不重启呢。
这里有个驱逐原则：（1）子集群中少节点的被驱逐
（2）节点号大的被驱逐
（3）负载高的节点被驱逐
我们中的就是第二条，OK，节点2重启来了，我们登陆系统，输出用户名/密码
3.定位故障原因
（1）查看操作系统日志
$ su - root
Password:
# tail -30f /var/log/messages
我又重新模拟了一遍，由于信息量很大，我从里面找出与网络有关的告警信息
Jul 17 20:05:25 rac2 avahi-daemon: Withdrawing address record for 192.168.2.102 on eth1.
收回eth1网卡的ip地址，导致节点1驱逐节点2，节点2自动重启
Jul 17 20:05:25 rac2 avahi-daemon: Leaving mDNS multicast group on interface eth1.IPv4 with address 192.168.2.102.
网卡eth1脱离多组播组
Jul 17 20:05:25 rac2 avahi-daemon: iface.c: interface_mdns_mcast_join() called but no local address available.
Jul 17 20:05:25 rac2 avahi-daemon: Interface eth1.IPv4 no longer relevant for mDNS.
网卡eth1不在与mDNS有关
Jul 17 20:09:54 rac2 logger: Oracle Cluster Ready Services starting up automatically.
Oracle集群自动启动
Jul 17 20:09:59 rac2 avahi-daemon: Registering new address record for fe80::20c:29ff:fe8f:f191 on eth1.
Jul 17 20:09:59 rac2 avahi-daemon: Registering new address record for 192.168.2.102 on eth1.
注册新ip地址
Jul 17 20:10:17 rac2 logger: Cluster Ready Services completed waiting on dependencies.
CRS完成等待依赖关系
从上面信息我们大体知道，是因为eth1网卡的问题导致节点2重启的，为了进一步分析问题我们还需要看一下CRS排错日志
# tail -100f $ORA_CRS_HOME/log/rac2/crsd/crsd.log
Abnormal termination by CSS, ret = 8
异常终止CSS
2013-07-17 20:11:18.115: [ default]0CRS Daemon Starting
2013-07-17 20:11:18.116: [ CRSMAIN]0Checking the OCR device
2013-07-17 20:11:18.303: [ CRSMAIN]0Connecting to the CSS Daemon
重启CRSCSS进程
# pwd
/u01/crs1020/log/rac2/cssd
# more ocssd.log 查看cssd进程日志
2013-07-17 17:26:18.319 >TRACE: clssgmclientlsnr: listening on (ADDRESS=(PROTOCOL=ipc)(KEY=OCSSD_LL_rac2_crs))
这里可以看到rac2节点的cssd进程监听出了问题
2013-07-17 17:26:19.296 >TRACE: clssnmHandleSync: Acknowledging sync: src srcName seq sync
请确认两个节点的同步问题
从以上一系列信息可以分析出这是内联网通信问题，由于两个节点的信息无法同步导致信息无法共享从而引起脑裂现象
4.节点2重启自动恢复正常状态
# ifconfig
eth0 Link encap:EthernetHWaddr 00:0C:29:8F:F1:87
inet addr:192.168.1.102Bcast:192.168.1.255Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe8f:f187/64 Scope:Link
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
RX packets:567 errors:0 dropped:0 overruns:0 frame:0
TX packets:901 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:65402 (63.8 KiB)TX bytes:96107 (93.8 KiB)
Interrupt:185 Base address:0x14a4
eth0:1 Link encap:EthernetHWaddr 00:0C:29:8F:F1:87
inet addr:192.168.1.202Bcast:192.168.1.255Mask:255.255.255.0
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
Interrupt:185 Base address:0x14a4
eth1 Link encap:EthernetHWaddr 00:0C:29:8F:F1:91
inet addr:192.168.2.102Bcast:192.168.2.255Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe8f:f191/64 Scope:Link
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
RX packets:76659 errors:0 dropped:0 overruns:0 frame:0
TX packets:51882 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:61625763 (58.7 MiB)TX bytes:26779167 (25.5 MiB)
Interrupt:193 Base address:0x1824
eth2 Link encap:EthernetHWaddr 00:0C:29:8F:F1:9B
inet addr:192.168.203.129Bcast:192.168.203.255Mask:255.255.255.0
inet6 addr: fe80::20c:29ff:fe8f:f19b/64 Scope:Link
UP BROADCAST RUNNING MULTICASTMTU:1500Metric:1
RX packets:409 errors:0 dropped:0 overruns:0 frame:0
TX packets:58 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:45226 (44.1 KiB)TX bytes:9567 (9.3 KiB)
Interrupt:169 Base address:0x18a4
lo Link encap:Local Loopback
inet addr:127.0.0.1Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNINGMTU:16436Metric:1
RX packets:49025 errors:0 dropped:0 overruns:0 frame:0
TX packets:49025 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:11292111 (10.7 MiB)TX bytes:11292111 (10.7 MiB)
我们看一下网卡ip地址，被收回的私有eth1网卡ip现在已经恢复了，这是因为刚刚节点2进行了重启操作。重启后会初始化所有网卡，被我们禁用的eth1网卡被重新启用，重新恢复ip。
检查CRS进程状态，全都是健康的
# crsctl check crs
CSS appears healthy
CRS appears healthy
EVM appears healthy
检查集群，实例，数据库，监听，ASM服务状态，也都是完好无损，全部启动了
# crs_stat -t
Name Type Target State Host
------------------------------------------------------------
ora....B1.inst application ONLINE ONLINE rac1
ora....B2.inst application ONLINE ONLINE rac2
ora....DB1.srv application ONLINE ONLINE rac1
ora.....TAF.cs application ONLINE ONLINE rac1
ora.RACDB.dbapplication ONLINE ONLINE rac1
ora....SM1.asmapplication ONLINE ONLINE rac1
ora....C1.lsnr application ONLINE ONLINE rac1
ora.rac1.gsd application ONLINE ONLINE rac1
ora.rac1.ons application ONLINE ONLINE rac1
ora.rac1.vip application ONLINE ONLINE rac1
ora....SM2.asmapplication ONLINE ONLINE rac2
ora....C2.lsnr application ONLINE ONLINE rac2
ora.rac2.gsd application ONLINE ONLINE rac2
ora.rac2.ons application ONLINE ONLINE rac2
ora.rac2.vip application ONLINE ONLINE rac2
RAC故障分析并解决的整个过程到此结束
三模拟OCR磁盘不可用时，RAC会出现什么现象?给出故障定位的整个过程
OCR磁盘：OCR磁盘中注册了RAC所有的资源信息，包含集群、数据库、实例、监听、服务、ASM、存储、网络等等，只有被OCR磁盘注册的资源才能被CRS集群管理，CRS进程就是按照OCR磁盘中记录的资源来管理的，在我们的运维过程中可能会发生OCR磁盘信息丢失的情况，例如在增减节点时，添加 or 删除OCR磁盘时可能都会发生。接下来我们模拟一下当OCR磁盘信息丢失时，如果定位故障并解决。
实验
1.检查OCR磁盘和CRS进程
（1）检查OCR磁盘，只有OCR磁盘没有问题，CRS进程才可以顺利管理
# ocrcheck
Status of Oracle Cluster Registry is as follows :
Version : 2
Total space (kbytes) : 104344
Used space (kbytes) : 4344
Available space (kbytes) : 100000
ID :1752469369
Device/File Name : /dev/raw/raw1 这个就是OCR磁盘所属的裸设备
Device/File integrity check succeeded
Device/File not configured
Cluster registry integrity check succeeded 完整检查完毕没有问题
（2）检查CRS状态
# crsctl check crs
CSS appears healthy
CRS appears healthy
EVM appears healthy
集群进程全部健康
（3）关闭CRS守护进程
# crsctl stop crs
Stopping resources. 停止资源
Successfully stopped CRS resources 停止CRS进程
Stopping CSSD. 停止CSSD进程
Shutting down CSS daemon.
Shutdown request successfully issued.
关闭请求执行成功
# crsctl check crs
Failure 1 contacting CSS daemon 连接CSS守护进程失败
Cannot communicate with CRS 无法与CRS通信
Cannot communicate with EVM 无法与EVM通信
2.用root用户导出OCR磁盘内容进行OCR备份
# ocrconfig -export /home/oracle/ocr.exp
$ pwd
/home/oracle
$ ll
total 108
-rw-r--r-- 1 root root 98074 Jul 18 11:20 ocr.exp 已经生成OCR导出文件
3.重启CRS守护进程
# crsctl start crs
Attempting to start CRS stack 尝试启动CRS
The CRS stack will be started shortly CRS即将启动
检查CRS状态
# crsctl check crs 很好，我们重新启动后就变正常了
CSS appears healthy
CRS appears healthy
EVM appears healthy
4.使用裸设备命令0字节覆盖OCR磁盘内容模拟丢失状态
# dd if=/dev/zero of=/dev/raw/raw1 bs=1024 count=102400
102400+0 records in 102400记录输入
102400+0 records out 102400记录输出
104857600 bytes (105 MB) copied, 76.7348 seconds, 1.4 MB/s
命令解释
dd：指定大小的块拷贝一个文件，并在拷贝的同时进行指定的转换
if=/dev/zero 指定源文件，0设备
of=/dev/raw/raw1 指定目标文件，OCR磁盘
bs=1024 指定块大小1024个字节，即1k
count=102400 指定拷贝的块数，102400个块
5.再次检查OCR磁盘状态
# ocrcheck
PROT-601: Failed to initialize ocrcheck 初始化OCR磁盘失败
检查CRS状态
# crsctl check crs
Failure 1 contacting CSS daemon 连接CSS守护进程失败
Cannot communicate with CRS 无法与CRS通信
EVM appears healthy
CRS进程失败很正常，你想想连记录的资源信息都丢失了，还怎么管理呢
6.使用import恢复OCR磁盘内容
# ocrconfig -import /home/oracle/ocr.exp
7.最后检查OCR磁盘状态
谢天谢地顺顺利利恢复回来了
# ocrcheck
Status of Oracle Cluster Registry is as follows :
Version : 2
Total space (kbytes) : 104344
Used space (kbytes) : 4348
Available space (kbytes) : 99996
ID :425383787
Device/File Name : /dev/raw/raw1
Device/File integrity check succeeded
Device/File not configured
Cluster registry integrity check succeeded
8.关注CRS守护进程
# crsctl check crs
CSS appears healthy
CRS appears healthy
EVM appears healthy
非常好，当OCR磁盘恢复之后自动重启CRS守护进程
# crs_stat -t
Name Type Target State Host
------------------------------------------------------------
ora....B1.inst application ONLINE ONLINE rac1
ora....B2.inst application ONLINE OFFLINE
ora....DB1.srv application ONLINE ONLINE rac1
ora.....TAF.cs application ONLINE ONLINE rac1
ora.RACDB.db application ONLINE ONLINE rac1
ora....SM1.asmapplication ONLINE ONLINE rac1
ora....C1.lsnr application ONLINE ONLINE rac1
ora.rac1.gsd application ONLINE ONLINE rac1
ora.rac1.ons application ONLINE ONLINE rac1
ora.rac1.vip application ONLINE ONLINE rac1
ora....SM2.asmapplication ONLINE OFFLINE
ora....C2.lsnr application ONLINE OFFLINE
ora.rac2.gsd application ONLINE OFFLINE
ora.rac2.ons application ONLINE OFFLINE
ora.rac2.vip application ONLINE ONLINE rac2
我重启了一遍CRS集群服务
# ./init.crs stop
Shutting down Oracle Cluster Ready Services (CRS):
Stopping resources.
Successfully stopped CRS resources
Stopping CSSD.
Shutting down CSS daemon.
Shutdown request successfully issued.
Shutdown has begun. The daemons should exit soon.
# crs_stat -t
CRS-0184: Cannot communicate with the CRS daemon.
# ./init.crs start
Startup will be queued to init within 90 seconds.
现在都恢复了
$ crs_stat -t
Name Type Target State Host
----------------------------------------------------------------------------------------------
ora....B1.inst application ONLINE ONLINE rac1
ora....B2.inst application ONLINE ONLINE rac2
ora....DB1.srv application ONLINE ONLINE rac2
ora.....TAF.cs application ONLINE ONLINE rac2
ora.RACDB.dbapplication ONLINE ONLINE rac2
ora....SM1.asmapplication ONLINE ONLINE rac1
ora....C1.lsnr application ONLINE ONLINE rac1
ora.rac1.gsd application ONLINE ONLINE rac1
ora.rac1.ons application ONLINE ONLINE rac1
ora.rac1.vip application ONLINE ONLINE rac1
ora....SM2.asmapplication ONLINE ONLINE rac2
ora....C2.lsnr application ONLINE ONLINE rac2
ora.rac2.gsd application ONLINE ONLINE rac2
ora.rac2.ons application ONLINE ONLINE rac2
ora.rac2.vip application ONLINE ONLINE rac2
四模拟votedisk不可用时，RAC会出现什么现象?给出故障定位的整个过程
表决磁盘：在发生脑裂问题时，通过表决磁盘来决定驱逐哪个节点。这是发生在集群层上的脑裂。
控制文件：如果是发生在实例层上的脑裂问题，通过控制文件来决定驱逐哪个节点。
Votedisk冗余策略：
（1）votedisk可以选择外部冗余，通过外部的机制进行保护
（2）votedisk还可以选择Oracle自己的内部冗余，通过添加votedisk磁盘镜像来实现内部冗余
实验
1.检查vote disk状态
$ crsctl query css votedisk
0.0/dev/raw/raw2显示2号裸设备为表决磁盘
located 1 votedisk(s).只定位1个表决磁盘
2.停止CRS集群
# crsctl stop crs
Stopping resources.停止资源
Successfully stopped CRS resources 停止CRS进程
Stopping CSSD.停止CSSD进程
Shutting down CSS daemon.
Shutdown request successfully issued.
3.添加votedisk表决磁盘，实现内部冗余，
crsctladdcssvotedisk /dev/raw/raw3 –force把raw3这块裸设备添加入表决磁盘组
添加之后Oracle就会把原来表决磁盘内容复制一份到新表决磁盘中
4.再次检查vote disk状态
crsctl query css votedisk
5.启动CRS集群
# crsctl start crs
Attempting to start CRS stack 尝试启动CRS
The CRS stack will be started shortlyCRS即将启动
小结：当表决磁盘/dev/raw/raw2损坏时，可以用其镜像/dev/raw/raw3来代替，使其RAC可以继续对外提供服务。
RAC ERROR OCR VOTEDISK 性能优化
Leonarding
2013.07.18
北京&summer
分享技术~成就梦想
Blog：www.leonarding.com
页:
[1]