11.2 RAC: In "crsctl stat res -t" State Details May Be Missing or Incorrect [ID 1086563.1]

本文探讨了在Oracle Server Enterprise Edition 11.2及后续版本中使用crsctl stat res-t命令时出现的状态详情缺失或错误的问题。具体表现为部分ASM和DB资源的状态详情不正确或完全缺失,即使资源实际上处于在线状态。文章详细介绍了两种常见情况及其原因,并指出了这些错误仅具信息性质,不影响实际功能。

Applies to:

Oracle Server - Enterprise Edition - Version: 11.2.0.1 and later   [Release: 11.2 and later ]
Information in this document applies to any platform.

Symptoms

In output of "crsctl stat res -t", for some or all ASM and/or DB resources, the state_details are missing or incorrect.
Case 1
The state_details may be blank instead of showing "Open" or "Started" when the resource is in fact open/started, although STATE correctly lists the resource as online:
 
$ crsctl stat res -t ------------------------------------------------------------- NAME TARGET STATE SERVER STATE_DETAILS ------------------------------------------------------------- Local Resources ------------------------------------------------------------- ora.asm ONLINE ONLINE oranode1 <<=== missing ONLINE ONLINE oranode2 Started ------------------------------------------------------------- Cluster Resources ------------------------------------------------------------- ora.tst.db 1 ONLINE ONLINE oranode1 <<=== missing 2 ONLINE ONLINE oranode2 Open

In this case, the STATE details only show for the most recently started instance of each group ie. in the above, +ASM2 and TST2 were the most recently started.
Case 2
After restarting CRS stack on one node, the state_details of some resources on the other node(s) may simply be wrong and not matching with the state, or may be missing altogether:
       


 
$ crsctl stat res -t ------------------------------------------------------------- NAME TARGET STATE SERVER STATE_DETAILS ------------------------------------------------------------- Local Resources ------------------------------------------------------------- ora.asm ONLINE ONLINE oranode1 Started ONLINE ONLINE oranode2 <<=== missing ------------------------------------------------------------- Cluster Resources ------------------------------------------------------------- ora.tst.db 1 ONLINE ONLINE oranode1 Open 2 ONLINE ONLINE oranode2 Instance Shutdown <<=== wrong

Cause

Case 1
This is caused by partially published bug:
BUG 9377625 - STATE_DETAILS ONLY SHOWED LAST STARTED ASM "STARTED" FOR 'CRSCTL STATUS RES -T"
This bug will be fixed in future release 12.1.

Case 2
This is caused by unpublished bug:
BUG 8516657 (Unpublished): CONFUSING 'STATE_DETAILS' FOR ASM/DB INSTANCE FROM 'CRSCTL STAT RES -T'

Solution

These issues will be patched in future versions and/or CRS PSU's.  Check My Oracle Support to see if one-off patches are available.
In the meantiume:
  • These errors are informational only, in other words, functionality of the database, asm or other resources are not affected.  Further, the STATE is not affected, only STATE details may be inaccurate.  Inaccurate STATE details can simply be ignored.
@INTERNAL NOTE:  unpublishing this piece because 'crs_stat' is officially deprecated in 11.2: 
@crs_stat -l and/or crs_stat -t can be used to give a listing without STATE details.

References

BUG:9377625 - STATE_DETAILS ONLY SHOWED LAST STARTED ASM "STARTED" FOR 'CRSCTL STATUS RES -T"
<think>好的,我现在需要帮助用户解决Oracle 11g RAC节点2上运行crsctl stat res -t命令无响应的问题。首先,我要理清用户的问题描述,然后根据提供的引用信息和自己的知识进行分析。 用户提到节点2执行crsctl stat res -t没有响应,而根据引用内容,类似的问题在引用2和引用4中都有涉及。例如,引用2提到state_details列可能显示空白,而STATE列正确显示online。引用4则显示crsctl stat res -t可能没有实例显示,但其他节点正常。这可能涉及资源状态同步或通信问题。 首先,我需要考虑可能的原因。常见的原因包括网络问题、集群软件(CRS)状态异常、资源锁冲突或节点间通信故障。另外,Oracle RAC的日志文件是排查问题的关键,比如检查CRSD日志、OHASD日志等。 接下来,按照步骤进行排查: 1. **检查集群通信状态**:用户需要确认节点间的网络连接是否正常,特别是私网和公网接口。使用oifcfg getif查看网卡配置是否正确,确保心跳网络正常。如果网络有问题,可能导致节点无法通信,CRS无法同步资源状态。 2. **验证集群服务状态**:在节点2上,检查CRS服务是否运行正常。使用crsctl check crs命令检查各个服务的状态,如CSSD、CRSD等。如果某个服务未运行,需要启动或重启相关服务。 3. **检查资源状态**:在正常节点(如节点1)运行crsctl stat res -t,查看节点2资源的状态。如果节点2的资源显示OFFLINE或UNKNOWN,可能需要手动干预,比如使用crsctl stop/start resource命令重启资源。 4. **查看日志文件**:定位到CRSD和OHASD的日志,通常在$GRID_HOME/log/<节点名>目录下。查看最近的错误或警告信息,如无法连接到其他节点、资源启动超时等。 5. **检查OCR和Voting Disk**:OCR和Voting Disk的损坏或访问问题会导致集群无法正确管理资源。使用ocrcheck和crsctl query css votedisk命令验证其完整性。如果有问题,可能需要恢复或替换。 6. **节点重启**:如果上述步骤无效,尝试重启节点2的CRS服务。使用crsctl stop crs和crsctl start crs命令。注意,在重启前确保不会影响生产环境。 在引用3中提到了一些管理命令,比如srvctl检查实例状态,关闭或启动数据库实例。用户可能需要使用这些命令来进一步诊断节点2的具体情况。 另外,引用2中提到的state_details列空白可能与资源未正确启动有关,需要确认资源是否处于预期状态。例如,数据库实例是否真的已经打开,或者存在依赖项未满足。 综合以上信息,解决方案可能包括修复网络问题、重启CRS服务、恢复OCR/Voting Disk,或手动调整资源状态。同时,用户需要根据日志中的具体错误信息进行针对性处理。</think>以下为Oracle 11g RAC节点2上`crsctl stat res -t`无响应的排查步骤及解决方案: ### 一、问题原因分析 1. **集群通信故障** 节点间心跳网络异常(如私网网卡故障、IP冲突)会导致CRS无法同步资源状态[^2][^3] 2. **CRS服务异常** 节点2的Cluster Ready Services未正常运行,常见症状: ```bash $ crsctl check crs CRS-4638: Oracle High Availability Services is online CRS-4535: Cannot communicate with Cluster Ready Services CRS-4529: Cluster Synchronization Services is online ``` 3. **资源锁冲突** 节点资源被异常锁定,常见于Voting Disk或OCR磁盘组访问冲突[^4] 4. **日志文件异常** CRSD日志(`$GRID_HOME/log/<节点名>/crsd/crsd.log`)可能包含I/O错误或超时记录 ### 二、分步解决方案 #### 步骤1:验证集群基础状态 ```bash # 检查节点间连通性(在节点1执行) $ cluvfy comp nodecon -n node2 # 查看网卡配置(节点2执行) $ oifcfg getif ``` #### 步骤2:检查CRS服务状态 ```bash # 查看服务状态(节点2执行) $ crsctl check crs # 若CRSD服务异常,强制重启 $ crsctl stop crs -f $ crsctl start crs ``` #### 步骤3:检查资源锁定状态 ```bash # 查看Voting Disk状态(任意节点执行) $ crsctl query css votedisk # 验证OCR健康度 $ ocrcheck ``` #### 步骤4:分析关键日志 ```bash # 查看最新CRSD日志(节点2执行) $ tail -500f $GRID_HOME/log/`hostname`/crsd/crsd.log | grep -E 'ERROR|WARNING' # 检查OHASD日志 $ tail -200 $GRID_HOME/log/`hostname`/ohasd/ohasd.log ``` ### 三、典型场景解决方案 | 场景 | 现象特征 | 解决方案 | |------|----------|----------| | 网络隔离 | 日志出现`CSSD: Failed to check IPC` | 修复私网连接,重启CRS | | OCR损坏 | `ocrcheck`报`PROT-602`错误 | 使用`ocrconfig -restore`恢复备份 | | 资源死锁 | 日志出现`CRS-0215`错误代码 | 执行`crsctl delete resource`强制释放资源 | ### 四、预防措施 1. 定期执行集群验证工具: ```bash $ cluvfy stage -post crsinst -n all ``` 2. 配置CRS自动重启: ```bash $ crsctl modify resource "ora.cssd" -attr AUTO_START=1 ```
评论
成就一亿技术人!
拼手气红包6.0元
还能输入1000个字符
 
红包 添加红包
表情包 插入表情
 条评论被折叠 查看
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值