linux 5.8 11.1.0.7 rac 绑定双网卡的public ip 拔掉网线vip不漂移...

makaveli

现在有套linux 5.8上的11.1.0.7的rac库，已经打到psu7.12，public ip为两块网卡绑定，现在进行故障切换测试，同时拔掉一台机器两根public网卡的网线后，vip没有进行切换，请各位帮忙诊断下是哪里的问题。

makaveli

问题解决了，谢谢2群的小江帮忙。
参考VIP does not Failover or Network Resource Stays Online after Disconnected Cables for Bond Public Network [ID 1276737.1]文档。通过mii-tool检查正常，参考文档解决 bug 10020138的步骤修改racgvip文件后vip能够正常漂移了。

1. Another potential solution is fix for bug 10020138 (fixed 11.2.0.3, 12.1). Due to this bug, interface check may not use mii-tool as _LINK_STAT was not reset. The workaround is to locate the following in $ORA_CRS_HOME/bin/racgvip:
   
2. 
   
3.        else
   
4.          logx "checkIf: ethtool checked if=$_IF failed"
   
5.        fi
   
6. 
   
7. And change it to:
   
8. 
   
9.        else
   
10.          logx "checkIf: ethtool checked if=$_IF failed"
    
11.          _LINK_STAT=
    
12.        fi

复制代码

Liu Maclean(刘相兵

就你的日志看 似乎有一次failover啊，14:58 12-19

1. 
   
2. 2012-12-18 14:36:32.587: [  CRSRES][1509714240] Start of `ora.nsbxdb2.vip` on member `nsbxdb1` succeeded.
   
3. 2012-12-18 14:38:05.116: [  OCRUTL][1295698240]u_freem: mem passed is null
   
4. 2012-12-18 14:38:06.572: [    CRSD][1497106752] SM: rE2Ec: 4
   
5. 2012-12-18 14:38:06.573: [  CRSRES][1536997696] StopResource: setting CLI values
   
6. 2012-12-18 14:38:06.955: [    CRSD][1536997696] SM:dE2Ec: all E2E cmds done. 0
   
7. 2012-12-19 14:58:27.154: [ CRSCOMM][1532795200] CLEANUP: Searching for connections to failed node nsbxdb2
   
8. 2012-12-19 14:58:27.154: [  CRSEVT][1532795200] Processing member leave for nsbxdb2, incarnation: 250787480
   
9. 2012-12-19 14:58:27.155: [    CRSD][1532795200] SM: recovery in process: 8
   
10. 2012-12-19 14:58:27.155: [  CRSEVT][1532795200] Do failover for: nsbxdb2
    
11. 2012-12-19 14:58:27.232: [  OCRSRV][1532795200]s_update_remote_cache_int: FAILED TO RCV ACK FROM node 2 retcode 7
    
12. 2012-12-19 14:58:27.292: [  CRSRES][1532795200]  startup = 0
    
13. 2012-12-19 14:58:27.301: [  CRSRES][1532795200]  startup = 0
    
14. 2012-12-19 14:58:27.307: [  CRSRES][1532795200]  startup = 0
    
15. 2012-12-19 14:58:27.314: [  CRSRES][1532795200]  startup = 0
    
16. 2012-12-19 14:58:27.320: [  CRSRES][1532795200]  startup = 0
    
17. 2012-12-19 14:58:27.327: [  CRSRES][1532795200]  startup = 0
    
18. 2012-12-19 14:58:27.369: [  CRSRES][1534896448] startRunnable: setting CLI values
    
19. 2012-12-19 14:58:27.373: [  CRSRES][1534896448] Attempting to start `ora.nsbxdb2.vip` on member `nsbxdb1`
    
20. 2012-12-19 14:58:30.686: [  CRSRES][1534896448] Start of `ora.nsbxdb2.vip` on member `nsbxdb1` succeeded.
    
21. 2012-12-19 14:58:30.700: [  CRSEVT][1532795200] Post recovery done evmd event for: nsbxdb2
    
22. 2012-12-19 14:58:30.700: [    CRSD][1532795200] SM: recoveryDone: 0
    
23. 2012-12-19 14:58:30.702: [  CRSEVT][1532795200] Processing RecoveryDone

复制代码

VIP does not Failover or Network Resource Stays Online after Disconnected Cables for Bond Public Network [ID 1276737.1]        To Bottom       
10g/11gR1: VIP does not relocate back to the home node even after the public network problem is resolved [ID 805969.1]        To Bottom       
11gR2 VIP Does not Fail Over After Public Network Issue if Application VIP is ONLINE and Added With "appvipcfg create" [ID 1457380.1]

makaveli

Liu Maclean(刘相兵 发表于 2012-12-20 14:52
就你的日志看 似乎有一次failover啊，14:58 12-19VIP does not Failover or Network Resource Stays Online ...

谢ml提供的文档，这个问题已经解决了。