hpux11.31下安装rac 出现的故障。

esailing

环境：HPUX11.31 ，oracle crs 10.2.0.1.0 ，安装两节点的集群，ocr与vote 采用裸设备，ocr使用两个裸设备，vote使用三个裸设备，对应的5个lv是从盘柜中的一个vg中新建的，主机到盘柜共有四条链路。

现象：
安装crs时，需要在两个节点上以root用户分别在两个节点上执行两个脚本的界面。
在节点1和节点2上执行第一个脚本正常，只有在节点2上在执行root.sh脚本时出现提示：

root@DB2:/ # /oracle/crs/root.sh
WARNING: directory '/oracle' is not owned by root
Checking to see if Oracle CRS stack is already configured

Setting the permissions on OCR backup directory
Setting up NS directories
PROT-1: Failed to initialize ocrconfig
Failed to upgrade Oracle Cluster Registry configuration



alertdb2.log 日志文件内容：

[client(6441)]CRS-1009:The OCR configuration is invalid. Details in /oracle/crs/log/db2/client/ocrconfig_6441.log.



ocrconfig_6441.log 日志文件内容：

more ocrconfig_6441.log
Oracle Database 10g CRS Release 10.2.0.1.0 Production Copyright 1996, 2005 Oracle.  All rights re
served.
2009-01-16 19:09:47.112: [ OCRCONF][1]ocrconfig starts...
2009-01-16 19:09:47.112: [ OCRCONF][1]Upgrading OCR data
2009-01-16 19:09:47.113: [  OCROSD][1]utread:3: problem reading buffer ffffb070 buflen 4096 retva
l -1 phy_offset 102400 retry 0
2009-01-16 19:09:47.113: [  OCROSD][1]utread:4: problem reading the buffer errno 5 errstring I/O
error
2009-01-16 19:09:47.114: [  OCROSD][1]utread:3: problem reading buffer ffffb070 buflen 4096 retva
l -1 phy_offset 102400 retry 0
2009-01-16 19:09:47.114: [  OCROSD][1]utread:4: problem reading the buffer errno 5 errstring I/O
error
2009-01-16 19:09:47.114: [  OCRRAW][1]proprioini: OCR configuration on disk 0 is not valid
2009-01-16 19:09:47.114: [  OCRRAW][1]proprinit: Could not open raw device
2009-01-16 19:09:47.114: [ default][1]a_init:7!: Backend init unsuccessful : [26]
2009-01-16 19:09:47.115: [ OCRCONF][1]Exporting OCR data to [OCRUPGRADEFILE]
2009-01-16 19:09:47.115: [  OCRAPI][1]a_init:7!: Backend init unsuccessful : [33]
2009-01-16 19:09:47.115: [ OCRCONF][1]There was no previous version of OCR. error:[PROC-33: Oracl
e Cluster Registry is not configured]
2009-01-16 19:09:47.115: [  OCROSD][1]utread:3: problem reading buffer ffffb070 buflen 4096 retva
l -1 phy_offset 102400 retry 0
2009-01-16 19:09:47.115: [  OCROSD][1]utread:4: problem reading the buffer errno 5 errstring I/O
error
2009-01-16 19:09:47.115: [  OCROSD][1]utread:3: problem reading buffer ffffb070 buflen 4096 retva
l -1 phy_offset 102400 retry 0
2009-01-16 19:09:47.115: [  OCROSD][1]utread:4: problem reading the buffer errno 5 errstring I/O
error
2009-01-16 19:09:47.116: [  OCRRAW][1]proprioini: OCR configuration on disk 0 is not valid
2009-01-16 19:09:47.116: [  OCRRAW][1]proprinit: Could not open raw device
2009-01-16 19:09:47.116: [ default][1]a_init:7!: Backend init unsuccessful : [26]
2009-01-16 19:09:47.116: [  OCROSD][1]utread:3: problem reading buffer ffffb070 buflen 4096 retva
l -1 phy_offset 102400 retry 0
2009-01-16 19:09:47.116: [  OCROSD][1]utread:4: problem reading the buffer errno 5 errstring I/O
error
2009-01-16 19:09:47.116: [  OCROSD][1]utread:3: problem reading buffer ffffb070 buflen 4096 retva
l -1 phy_offset 102400 retry 0
2009-01-16 19:09:47.116: [  OCROSD][1]utread:4: problem reading the buffer errno 5 errstring I/O
error
2009-01-16 19:09:47.116: [  OCRRAW][1]proprioini: OCR configuration on disk 0 is not valid
2009-01-16 19:09:47.116: [  OCRRAW][1]proprinit: Could not open raw device
2009-01-16 19:09:47.116: [  OCRAPI][1]a_init:6b!: Backend init unsuccessful : [26]
2009-01-16 19:09:47.116: [ OCRCONF][1]Failed to initialized OCR context. error:[PROC-26: Error wh
ile accessing the physical storage Operating System error [I/O error] [5]]
2009-01-16 19:09:47.116: [ OCRCONF][1]Exiting [status=failed]...


以下是节点DB1上的现状：

在节点DB1上是正常的：

root@DB1:/oracle/crs # /oracle/crs/root.sh
WARNING: directory '/oracle' is not owned by root
Checking to see if Oracle CRS stack is already configured

Setting the permissions on OCR backup directory
Setting up NS directories
Oracle Cluster Registry configuration upgraded successfully
WARNING: directory '/oracle' is not owned by root
Successfully accumulated necessary OCR keys.
Using ports: CSS=49895 CRS=49896 EVMC=49898 and EVMR=49897.
node <nodenumber>: <nodename> <private interconnect name> <hostname>
node 1: db1 db1-priv db1
node 2: db2 db2-priv db2
Creating OCR keys for user 'root', privgrp 'sys'..
Operation successful.
Now formatting voting device: /dev/vg_ocr_vote/rvote01
Now formatting voting device: /dev/vg_ocr_vote/rvote02
Now formatting voting device: /dev/vg_ocr_vote/rvote03
Format of 3 voting devices complete.
Startup will be queued to init within 30 seconds.
Adding daemons to inittab
Expecting the CRS daemons to be up within 600 seconds.
CSS is active on these nodes.
        db1
CSS is inactive on these nodes.
        db2
Local node checking complete.
Run root.sh on remaining nodes to start CRS daemons.


表决盘的信息：
./crsctl query css votedisk
0.     0    /dev/vg_ocr_vote/rvote01
1.     0    /dev/vg_ocr_vote/rvote02
2.     0    /dev/vg_ocr_vote/rvote03


ocr的信息：

crcheck./
Status of Oracle Cluster Registry is as follows :
         Version                  :          2
         Total space (kbytes)     :     356264
         Used space (kbytes)      :        308
         Available space (kbytes) :     355956
         ID                       : 1971749425
         Device/File Name         : /dev/vg_ocr_vote/rocr01
                                    Device/File integrity check succeeded
         Device/File Name         : /dev/vg_ocr_vote/rocr02
                                    Device/File integrity check succeeded

         Cluster registry integrity check succeeded

[ 本帖最后由 esailing 于 2009-1-18 10:25 编辑 ]

esailing

请教下大家，这个现象是什么原因？如何解决啊？

esailing

卸载crs后，删除了相关lv和vg，pvcreate了相应的pv，在vgcreate时对指向物理卷的多路径进行了pvlink，建好lv后，再次安装crs，在需要执行两个脚本的时候，在两个节点上执行第一个脚本，都没有问题，让我郁闷的是，在节点1和节点2上分别执行第二个脚本root.sh,都出现了提示信息：

/oracle/crs/root.sh
WARNING: directory '/oracle' is not owned by root
Checking to see if Oracle CRS stack is already configured

Setting the permissions on OCR backup directory
Setting up NS directories
PROT-1: Failed to initialize ocrconfig
Failed to upgrade Oracle Cluster Registry configuration

ronnyzong

你提到了用了PVlink，这个应该是错误的做法，在11.31里LVM已经集成了多路径管理功能，不需要再加额外的Pvlink来做多路径的管理，具体用法可参考11。31的逻辑卷管理一书。
另，发生你这样的问题，主要从两个方面去做分析
1，检查日志：
$ORA_CRS_HOME/log/client/logs
看里面有没有什么提示
2，检查另一个日志文件：
ocrconfig.log

有没有类似以下的描述：
/u02/app/oracle/oradata/crs/clusterregistry, ret -1, errno 75, os err string Value too large for defined data type 2007-10-30 11:23:52.101: [ OCROSD][3085960896]utopen:6'': OCR location


解决方法：
1，检查系统是否有NFS文件系统，如有，检查Mount选项及权限是否正确
2，检查CRS相关的裸设备及文件系统的权限及属主是否正确

做完以下工作后再试一次：）

Good Luck

esailing

感谢楼上的回复，我现在不在用户现场，没有环境可以测试。能告诉我你的联系方式吗？我好直接请教您。