Oracle RAC Common installation, management errors

1. Install CRS failure, or perform root.sh error, possible causes:

　　(1) Time between nodes is out of sync, workaround: Use NTP service (2) The default firewall is enabled under Linux, resulting in execution of root.sh error: Failure at final check of the Oracle CRS stack.　　10 Workaround: Disable iptables, comment/etc/pam.d/other; # service Iptables stop;　　# chkconfig iptables off. (3) Permissions issues for bare devices, possibly due to changes in permissions after the operating system restarts. (RHEL4) Workaround: Add chown oracle:dba/dev/raw/raw* command to/etc/rc.local, automatically execute or modify file on each boot/etc/udev/permissions.d/50-　　Udev.permissions line 113th raw/*:root:disk:0660 to raw/*:oracle:dba:0660 (4) Solaris uses a disk partition that includes cylinder 0 to store OCR or vote disk.　　Workaround: The relevant partition should not include cylinder 0, which can start from 1.　　(5) The use of the public network IP address is not routable, workaround: Add the relevant Gateway (6) in the/etc/hosts no loopback address, that is 127.0.0.1 localhost (7) host name contains large letters, minus or underscore the special characters;　　(8) Oracle in HPUX do not use the GNU bash to modify the default shell;　　(9) Check whether the operating system, third-party cluster is officially supported by Oracle, whether the need for patches, such as the installation of Oracle 10g/11g Rac,oslevel on AIX5.3+HACMP requires 06 and above;　　() Aix platform, the Reserve_policy (Reserve_lock) attribute of the shared device needs to be modified to No_reserve (no);　　(11) All nodes see the OCR and vote device path name should be consistent, if not consistent, can be resolved with a soft connection;　　(12) Heartbeat device problem or ocr/votedisk access problem, unix/linux see if there is/tmp/crsctl.* file, get error message; (13) Reinstall failed solution in the old installation environment of CRS: DD Clears OCR and vote disk, and uses the following languageThe old CRS configuration file rm-rf/usr/tmp/.oracle/var/tmp/.oracle/tmp/.oracle/etc/oracle/*/var/opt/oracle/* RM-RF/ETC/INIT.CSSD /etc/init.crs*/etc/init.evmd/etc/init.d/init.cssd/etc/init.d/init.crs rm-rf/etc/init.d/init.crsd/etc/init.d/ Init.evmd/etc/rc3.d/k96init.crs/etc/rc3.d/s96init.crs rm-rf/etc/rc.d/rc2.d/k96init.crs/etc/rc.d/rc2.d/ S96init.crs 2 Client sometimes error: Ora-12545:connect failed because target host or object does not exist ORA-12545: because the target host or object does not exist,　　Connection Failure Workaround: Set Local_listener initialization Parameter 3 if the Select node interface does not come out. (1)　　The HACMP environment needs to check that the Oracle user must be in the Hagsuser group.　　(2) If it is hacmp5.4, need to hit Oracle patch 6718715; (3) You can use the cluster profile cluster configuration file, the content template is as follows: Mycluster rac01 rac01-priv rac01-vip rac02 rac02-priv rac02-vip rac03 Rac03-priv rac03-vip RAC04 rac04-priv rac04-vip 4. AIX on database start error ora-27504 IPC Error creating OSD context ora-27300 OS system dependent operation:sendmsg failed with status: ora-27301 OS failure Message:message too long ora-27302 failure occurred at:sskgxpsnd1 reason: Network parameter not set UDP_recvspace/udp_sendspace 5. Windows platform, ORA-600 [Kccsbck_first] Workaround: Turn off Media Sense 6.　　System Cycle Restart: May be CRS caused, if because of CRS, first set CRSCTL disable CRS to prohibit the automatic start of Oracle CRS.　　Check the OS, CRSD, and CSSD logs to see if there is a CRS file (ls-lrt/tmp/crsctl*) under/tmp/to determine the reason for the CRS failure. 7. The data instance of the second node cannot mount, hang or error, cause 1: Use vendor clusterware, libskgxn2.so file link error, Workaround: Compare two node oracle_home/lib/ Libskgxn2 and crs_home/lib/libskgxn2* are the same, if different need to link cause 2: Any platform Oracle 9i, no network parameters set the UDP parameter caused udp_sendspace or UDP_ Recvspace less than db_block_size * Db_file_multiblock_read_count workaround: Set corresponding parameters, such as set on AIX udp_recvspace = 65536 Udp_sendspace = 65536 cause 3:aix/hacmp/oracle9i, the service IP workaround is defined in HACMP: cluster_interconnects reason 4 is defined in the initialization parameter: any platform, the wrong Cluster_ is set Interconnects workaround: Check and correct this parameter, 8.　　No bare devices can be identified when building the library;　　Cause 1:oracle,10.2.0.3, many platforms (such as AIX and Linux) have rawutl related bugs, the workaround: Restore the Rawutl tool in 10.2.0.1, which is in the $ORACLE _home/bin directory. Reason 2:oracle9i,aix platform, need to set environment variable export Pgsd_subsys=grpsvcs 9. EVM Resource automatic error Oac_init:2: Could not connect to server, CLSC Retcode = 9 Workaround: Turn off UDP ICMP rejections/etc/rc.d/inIt.d/iptables stop; Chkconfig iptables off