Oracle监听出现假死-多次出现过

yjh625

最近生产环境的数据库又出现假死的情况了，具体情况如下：

1、Oracle的版本信息如下;
SQL*Plus: Release 10.2.0.1.0 - Production on Fri Jul 20 23:00:26 2012

Copyright (c) 1982, 2005, Oracle.  All rights reserved.


Connected to:
Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - 64bit Production
With the Partitioning, OLAP and Data Mining options

2、查看css.log报错信息如下，发现每次监听假死都是这个错误：

2012-07-20 22:34:51.948: [  OCROSD][978922688]utgdv:2:ocr loc file  cannot  be opened
2012-07-20 22:34:51.948: [  OCROSD][978922688]utopen:1: Couldnt find ocr,[ocrmirror] location in config file
2012-07-20 22:34:51.948: [  OCRRAW][978922688]proprinit: Could not open raw device
2012-07-20 22:34:51.948: [ default][978922688]a_init:7!: Backend init unsuccessful : [33]
2012-07-20 22:34:51.948: [ CSSCLNT][978922688]clsssinit: error(33 ) in OCR initialization

3、listener.log的日志详见附件，期间处理措施为杀死监听，然后咋重启监听。

解决方法是知道怎么解决，但是总是出现监听假死，请刘大帮忙查找根本原因，谢谢。

Maclean Liu(刘相兵

没有明确的日志信息说明监听假死的原因， 但实在不推荐继续使用10.2.0.1这个版本

yjh625

1、系统内存情况
[kavin@xianjkoracle89 ~]$ free -m
             total       used       free     shared    buffers     cached
Mem:          2005       1989         15          0         19       1742
-/+ buffers/cache:        228       1777
Swap:        80003        316      79687
[kavin@xianjkoracle89 ~]$

2、网络是不存在问题的，网络我这边是一直有监控的

按照刘大的意思，这个是不是Oracle版本本身的bug导致的问题出现的？

Maclean Liu(刘相兵

如果仅仅讨论 listener重启的问题的话 ：

在 listener.log中没有找到太多关于监听假死的有用信息

1. 系统的free memory有多少？
2.

TNSLSNR for Linux: Version 10.2.0.1.0 - Production on 20-JUL-2012 22:34:51

Copyright (c) 1991, 2005, Oracle.  All rights reserved.

System parameter file is /app/soft/oracle/product/10.2.0/network/admin/listener.ora
Log messages written to /app/soft/oracle/product/10.2.0/network/log/listener.log
Trace information written to /app/soft/oracle/product/10.2.0/network/trace/listener.trc
Trace level is currently 0

Started with pid=23511
Listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=xianjkoracle89)(PORT=1521)))
Listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=ipc)(KEY=EXTPROC0)))
Listener completed notification to CRS on start

TIMESTAMP * CONNECT DATA [* PROTOCOL INFO] * EVENT [* SID] * RETURN CODE


是否确认 xianjkoracle89 这个host一直可以成功ping 是否存在网络问题


3.  10.2.0.1 这个版本实在不适合做 产品系统 ，建议你升级到10.2.0.5

teapot

10g/11gR1 Non-Cluster CRS: "PROC-33: Oracle Cluster Registry is not configured Operating System error" [ID 798603.1]

Applies to:
Oracle Server - Enterprise Edition - Version: 10.2.0.3 to 10.2.0.4 - Release: 10.2 to 10.2
This problem can occur on any platform.

Symptoms
Single node database get OCR errors in client log.
CSS & ASM are not in use.
ocr.loc is does not exist in /etc/oracle or /var/opt/oracle/


$ORACLE_HOME/log/hostname/client/
tail -200 clsc49.log

Oracle Database 10g CRS Release 10.2.0.1.0 Production Copyright 1996, 2005 Oracle. All rights reserved.
2009-01-28 13:17:53.515: [ OCROSD][1]utgdv:2:ocr loc file cannot be opened
2009-01-28 13:17:53.515: [ OCROSD][1]utopen:1: Couldnt find ocr,[ocrmirror] location in config file
2009-01-28 13:17:53.515: [ OCRRAW][1]proprinit: Could not open raw device
2009-01-28 13:17:53.516: [ default][1]a_init:7!: Backend init unsuccessful : [33]
2009-01-28 13:17:53.516: [ CSSCLNT][1]clsssinit: error(33 ) in OCR initialization


cat ocrconfig_9137.log

Oracle Database 10g CRS Release 10.2.0.1.0 Production Copyright 1996, 2005 Oracle. All rights reserved.
2008-12-09 16:30:57.153: [ OCRCONF][1]ocrconfig starts...
2008-12-09 16:30:57.172: [ OCRCONF][1]Failure initializing ocr in DEFAULT. REBOOT INSTALL. err :[PROC-32: Cluster Ready Services on the local node is not running Messaging error [9]]
2008-12-09 16:30:57.174: [ OCROSD][1]utgdv:2:ocr loc file cannot be opened
2008-12-09 16:30:57.174: [ OCROSD][1]utopen:1: Couldnt find ocr,[ocrmirror] location in config file
2008-12-09 16:30:57.174: [ OCRRAW][1]proprinit: Could not open raw device
2008-12-09 16:30:57.174: [ default][1]a_init:7!: Backend init unsuccessful : [33]
2008-12-09 16:30:57.174: [ OCRCONF][1]Failure in initializing ocr in INSTALL level. error:[PROC-33: Oracle Cluster Registry is not configured Operating System error [No such file or directory] [2]]
2008-12-09 16:30:57.175: [ OCRCONF][1]Exiting [status=failed]...


Cause
These errors are caused by
Bug 5938326 - CLUSTER STACK ERRORS THROWN IN NON-CLUSTER ENVIRONMENT

Some activity like gathering statistics or using v$asm views can result in those unnecessary logs and tracefiles. They occur because the database checks to see if it can talk to ASM. This check cannot be disabled after the first discovery which fails.

So when ASM is not in use and CSS is not installed and ASM queries are executed, excessive logs in trace files complaining about OCR not ready and/or CSS not running occur.


Solution
The bug is to be fixed in 10.2.0.5 and 11.1

The errors are of no consequence and can be ignored.
Workaround is to install CSS and enable CSS to run even if ASM is no in use (see <<Note.745271.1>>)

References
BUG:5938326 - CLUSTER STACK ERRORS THROWN IN NON-CLUSTER ENVIRONMENT
NOTE:745271.1 - How To Reconfigure the Cluster Synchronization Service (CSS) on Non-RAC & Trace It (Linux/Unix)?

[ 本帖最后由 teapot 于 2012-7-21 17:53 编辑 ]

aks5201314

楼主的数据库版本过旧，建议对数据库做一下升级。10.2.0.1问题还是不少的