HPUX11.31下安装11GRAC遇到的问题

1、[INS-40718] Single Client Access Name (SCAN):bydb-cluster-scan could not be resolved.
Cause - The name you provided as the SCAN could not be resolved using TCP/IP host name lookup. 
Action - Provide name to use for the SCAN for which the domain can be resolved.

这是因为/etc/hosts配置的scan和界面中指定的scan name不一致引起
hosts文件配置:
192.168.201.16   rac-scan.localdomain rac-scan
所以界面中的SCAN Name要是rac-scan才不会报错。

2、incorrect setting for minor number
在安装oracle 11g rac的时候会报错如下：
Checking settings of device file "/dev/async"
  Node Name     Available                 Comment                
  ------------  ------------------------  ------------------------
  racdb2        yes                       failed (incorrect setting for minor number.)
  racdb1        yes                       failed (incorrect setting for minor number.)
查了下support，得知：
Minor number   Description
0x0    This is the HP-UX default value for /dev/async
0x4    Enable disc device timeouts to complete with an error code rather than retrying forever. This setting is necessary                for application-level disc mirroring, so as to avoid the situation where the application waits forever for a failed disc            device to be repaired. Oracle RDBMS users should enable this feature when Automatic Storage Management      mirroring/replication (internal redundancy) is used. SGA will be locked in memory. 

0x100    Enable on-demand locking of memory pages by async driver when asyncdsk_open(2) is called. A low-overhead      routine is then used to lock a page into memory during I/O operations.
    On-demand locking is critically important when using Oracle's Automatic Memory Management feature (the use of
    MEMORY_TARGET
    in the
    init.ora
    file to control memory usage). RDBMS deployments utilizing dynamic nPar or dynamic vPar features should also      configure on-demand locking.
    More traditional RDBMS deployments can consider on-demand locking in light of its more obvious effects.      Generally speaking, RDBMS startup will be quicker because the complete SGA is not locked into memory       immediately. However, some instances will experience a slight run-time performance penalty with on-demand      locking as memory pages are dynamically locked/unlocked for each I/O request.
    
0x104     This is a combination of 0x100 and 0x4. Both the features are enabled
显然解决办法：
rm /dev/async
/sbin/mknod /dev/async c 101 0x104
chmod 666 /dev/async

异步io    ( http://blog.csdn.net/renfengjun/article/details/7972065)  其他人的解答
两个节点都需要设置，不然检查过不去。
XXJ-02#[/]ll /dev/async
crw-rw-rw-  1 bin        bin        101 0x000000 Aug 28 10:42 /dev/async
需要设置
XXJ-02#[/]rm /dev/async
XXJ-02#[/] /sbin/mknod /dev/async c 101 0x4 
XXJ-02#[/]chmod 666 /dev/async
XXJ-02#[/]ll /dev/async
crw-rw-rw-  1 root       sys        101 0x000004 Aug 28 14:14 /dev/async
如果这里没有发现设备，可以直接使用mknod命令创建

./runcluvfy.sh stage -pre crsinst -n racdb1,racdb2 -fixup -verbose
Checking settings of device file "/dev/async"
  Node Name     Available                 Comment                
  ------------  ------------------------  ------------------------
  racdb2        yes                       passed                 
  racdb1        yes                       passed                 
Result: Check for settings of device file "/dev/async" passed.

3、Check: User equivalence for user "grid"
  Node Name                             Status                 
  ------------------------------------  ------------------------
  bydb2                                 passed                 
  bydb1                                 failed                 
Result: PRVF-4007 : User equivalence check failed for user "grid"
将/home/grid/.ssh/known_hosts 清空，解决

参考故障：
linux中ssh登录Permanently added (RSA) to the list of known&nb
[root@www]# ssh admin@xxx.xxx.xxx.xx
2 The authenticity of host 'xxx.xxx.xxx.xxx' can't be established.
3 RSA key fingerprint is f4:d0:1a:9c:09:65:78:71:25:33:79:c2:d5:aa:0b:c1.
4 Are you sure you want to continue connecting (yes/no)? yes
5 Warning: Permanently added 'xxx.xxx.xxx.xxx' (RSA) to the list of known hosts.
6 Permission denied (publickey,gssapi-with-mic,password).
1）、可以把~/.ssh/known_hosts清除
2）、如果还是不行
修 改/etc/ssh/sshd-config文件，将其中的PermitRootLogin no修改为yes，PubkeyAuthentication yes修改为no，AuthorizedKeysFile .ssh/authorized_keys前面加上#屏蔽掉，PasswordAuthentication no修改为yes就可以了。
同时设置 UseDNS no，最后保存修改，重启sshd服务再连接这个问题就可以解决了。等同于第四步注释掉DNS
3）、如果还是有问题，那可以用下面的方式解决：
用命令 ssh -l username hostname
 1）、这个是ssh安全认证是的一个RSA认证。此处必须选择yes才能连接。第一次yes后，他会询问你是否永久把这个RSA认证加入本地，选择yes后，以后不会再出现提醒。每次登陆只需要输入密码即可。
 2）、也可以不用输入1中的yes，但是需要修改本机配置。
4）有能配置了DNS，没有配的话，没有下面两个文件。
在这里两个文件里注释掉相关信息
linux
/etc/ nsswitch.conf
/etc/resolv.conf
HP-UX 配置DNS服务器参数文件：
/etc/ nsswitch.conf
/etc/resolv.conf
 AIX配置DNS服务器参数文件
/etc/netsvc.conf
/etc/resolv.conf          ##如没有可以手动创建
修改server上/etc/nsswitch.conf中hosts为hosts： files

其他参考：
http://mohamedelazab.blogspot.in/2011/01/cluster-verify-utility-fails-with-user.html
http://csayantan.wordpress.com/11gr2-rac-installation/way-to-avoiding-prv-4007-in-runcluvfy-sh/
 

4、配置SSH发现不成功，经检查发现可能是文件权限问题。
修改/home/oracle的权限为0755。
然后修改~/.ssh/的权限为0700
修改其他文件的权限如下：
$ ls -l ~/.ssh
total 64
-rw-r--r--   1 oracle     oinstall       788 Jun 21 11:44 authorized_keys
-rw-------   1 oracle     oinstall      1675 Jun 21 11:40 id_rsa
-rw-r--r--   1 oracle     oinstall       394 Jun 21 11:40 id_rsa.pub
-rw-r--r--   1 oracle     oinstall      1326 Jun 21 11:39 known_hosts
然后测试成功。

5、如何处理11gR2 RAC下oc4j和gsd服务为OFFLINE状态？
http://blog.163.com/xiaohan_008/blog/static/126347420129994358263/
$ crs_stat -t
Name           Type           Target    State     Host       
------------------------------------------------------------
ora.CRSDATA.dg ora....up.type ONLINE    ONLINE    bydb1      
ora.DATA01.dg  ora....up.type ONLINE    ONLINE    bydb1      
ora....CKUP.dg ora....up.type ONLINE    ONLINE    bydb1      
ora....ER.lsnr ora....er.type ONLINE    ONLINE    bydb1      
ora....N1.lsnr ora....er.type ONLINE    ONLINE    bydb1      
ora.asm        ora.asm.type   ONLINE    ONLINE    bydb1      
ora.bydb.db    ora....se.type OFFLINE   OFFLINE              
ora....SM1.asm application    ONLINE    ONLINE    bydb1      
ora....B1.lsnr application    ONLINE    ONLINE    bydb1      
ora.bydb1.gsd  application    OFFLINE   OFFLINE              
ora.bydb1.ons  application    ONLINE    ONLINE    bydb1      
ora.bydb1.vip  ora....t1.type ONLINE    ONLINE    bydb1      
ora....SM2.asm application    ONLINE    ONLINE    bydb2      
ora....B2.lsnr application    OFFLINE   OFFLINE              
ora.bydb2.gsd  application    OFFLINE   OFFLINE              
ora.bydb2.ons  application    ONLINE    ONLINE    bydb2      
ora.bydb2.vip  ora....t1.type ONLINE    ONLINE    bydb2      
ora.cvu        ora.cvu.type   ONLINE    ONLINE    bydb1      
ora.gsd        ora.gsd.type   OFFLINE   OFFLINE              
ora....network ora....rk.type ONLINE    ONLINE    bydb1      
ora.oc4j       ora.oc4j.type  OFFLINE   OFFLINE              
ora.ons        ora.ons.type   ONLINE    ONLINE    bydb1      
ora.scan1.vip  ora....ip.type ONLINE    ONLINE    bydb1      
$ lsnrctl start
那么，是我们的RAC配置有问题吗?这两类服务没有ONLINE会不会影响RAC数据库呢？我们又该如何启动这些服务呢？
        其实，我们的RAC配置是没有问题的，也不会影响数据库正常运行。在默认情况下，其中的ora.gsd服务状态是OFFLINE状态，我们可不予理会，除非需要在11g Cluster上配置Oracle 9iR2的RAC。但凡以grid用户执行crs_stat -t命令，看到Target和State这两列的状态均为OFFLINE，我们均可忽略。但是，我们该如何使之启动呢？
 1、 启动ora.oc4j服务：
$ srvctl start oc4j 
OC4J could not be started
PRCR-1079 : Failed to start resource ora.oc4j
CRS-2501: Resource 'ora.oc4j' is disabled
$ srvctl enable oc4j -h
Enable OC4J instance for Oracle Clusterware management.
Usage: srvctl enable oc4j [-n ] [-v]
    -n            Node name
    -v                       Verbose output
    -h                       Print usage
$ srvctl enable oc4j   
$ srvctl start oc4j -v
OC4J has been started

2、 启动ora.gsd服务：
$ srvctl status nodeapps 
VIP bydb1_vip is enabled
VIP bydb1_vip is running on node: bydb1
VIP bydb2_vip is enabled
VIP bydb2_vip is running on node: bydb2
Network is enabled
Network is running on node: bydb1
Network is running on node: bydb2
GSD is disabled
GSD is not running on node: bydb1
GSD is not running on node: bydb2
ONS is enabled
ONS daemon is running on node: bydb1
ONS daemon is running on node: bydb2
$ srvctl enable nodeapps -g
$ srvctl status nodeapps 
VIP bydb1_vip is enabled
VIP bydb1_vip is running on node: bydb1
VIP bydb2_vip is enabled
VIP bydb2_vip is running on node: bydb2
Network is enabled
Network is running on node: bydb1
Network is running on node: bydb2
GSD is enabled
GSD is not running on node: bydb1
GSD is not running on node: bydb2
ONS is enabled
ONS daemon is running on node: bydb1
ONS daemon is running on node: bydb2

然后，启动GSD服务 --------------因为没有配置GNS所以不能启动
$ srvctl start nodeapps -g
PRCR-1079 : Failed to start resource ora.gsd
CRS-5016: Process "/oracle/ora11g/crs_home/bin/gsdctl" spawned by agent "/oracle/ora11g/crs_home/bin/oraagent.bin" for action "start" failed: details at "(:CLSN00010:)" in "/oracle/ora11g/crs_home/log/bydb1/agent/crsd/oraagent_grid/oraagent_grid.log"
CRS-5016: Process "/oracle/ora11g/crs_home/bin/gsdctl" spawned by agent "/oracle/ora11g/crs_home/bin/oraagent.bin" for action "start" failed: details at "(:CLSN00010:)" in "/oracle/ora11g/crs_home/log/bydb2/agent/crsd/oraagent_grid/oraagent_grid.log"
CRS-2674: Start of 'ora.gsd' on 'bydb1' failed
CRS-2674: Start of 'ora.gsd' on 'bydb2' failed

$ srvctl start nodeapps
PRCR-1079 : Failed to start resource ora.gsd
CRS-5016: Process "/oracle/ora11g/crs_home/bin/gsdctl" spawned by agent "/oracle/ora11g/crs_home/bin/oraagent.bin" for action "start" failed: details at "(:CLSN00010:)" in "/oracle/ora11g/crs_home/log/bydb2/agent/crsd/oraagent_grid/oraagent_grid.log"
CRS-5016: Process "/oracle/ora11g/crs_home/bin/gsdctl" spawned by agent "/oracle/ora11g/crs_home/bin/oraagent.bin" for action "start" failed: details at "(:CLSN00010:)" in "/oracle/ora11g/crs_home/log/bydb1/agent/crsd/oraagent_grid/oraagent_grid.log"
CRS-2674: Start of 'ora.gsd' on 'bydb2' failed
CRS-2674: Start of 'ora.gsd' on 'bydb1' failed
PRKO-2421 : Network resource is already started on node(s): bydb1,bydb2
PRKO-2420 : VIP is already started on node(s): bydb1
PRKO-2420 : VIP is already started on node(s): bydb2
PRKO-2422 : ONS is already started on node(s): bydb1,bydb2