OBIEE-----ClusterControler通信的问题

　　今天，OBIEE11G做维护，重启了下。确出现了些问题。ClusterControler 无法通信的问题。本人觉得这个问题很有借鉴意义，希望与大家分享一下。

   1.重启后

   执行{Oracle_bieehome}/user_projects/domains/bifoundation_domain/bin/startweblogic.cmd 启动weblogic应用服务器

  执行{Oracle_bieehome}/user_projects/domains/bifoundation_domain/bin/startManageredWeblogic.cmd 启动weblogic守护进程。

  执行{Oracle_bieehome}/instance/instance1/bin/opmnctl startproc  ias-component=coreapplication_obiccs1

  执行{Oracle_bieehome}/instance/instance1/bin/opmnctl startproc ias-component=coreapplication_obisch1

  执行{Oracle_bieehome}/instance/instance1/bin/opmnctl startproc ias-component=coreapplication_obijh1

  执行{Oracle_bieehome}/instance/instance1/bin/opmnctl startproc ias-component=coreapplication_obips1

  执行{Oracle_bieehome}/instance/instance1/bin/opmnctl startproc ias-component=coreapplication_obis1

其中除了ClusterControler启动失败以外，其他BI BackGroud Program Process 均启动成功。

 2.重启 ClusterControler 是否能解决。

 为了验证ClusterControler的启动状态，

使用 {Oracle_bieehome}/instance/instance1/bin/opmnctl status

同时使用 http://biee_host:7001/em 进入coreapplication 中查看可行性状态，显示ClusterController状态为Failed

  单独启动ClusterController进程 restart ,启动依旧失败

  全部重启BIEE后台进程(五大后台进程，有依存关系) ，ClusterController依旧失败

 3.定位 ClusterController进程日志 确定问题

 opmn的日志文件位于  {Oracle_bieehome}/instance/instance1/diagnostics/logs/opmn/opmn中的

 

[2012-09-28T13:23:51+08:00] [opmn] [NOTIFICATION:1] [663] [pm-process] Stopping Process: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161224:2212)
[2012-09-28T13:23:51+08:00] [opmn] [ERROR:1] [] [libopmncustom] Forcefully Terminating Process: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161224:2212)
[2012-09-28T13:25:52+08:00] [opmn] [NOTIFICATION:1] [663] [pm-process] Stopping Process: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161224:2212)
[2012-09-28T13:25:52+08:00] [opmn] [ERROR:1] [] [libopmncustom] Forcefully Terminating Process: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161224:2212)
[2012-09-28T13:27:54+08:00] [opmn] [NOTIFICATION:1] [663] [pm-process] Stopping Process: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161224:2212)
[2012-09-28T13:27:54+08:00] [opmn] [ERROR:1] [] [libopmncustom] Forcefully Terminating Process: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161224:2212)
[2012-09-28T13:28:03+08:00] [opmn] [NOTIFICATION:1] [662] [pm-process] Starting Process: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161240:0)
[2012-09-28T13:28:05+08:00] [opmn] [NOTIFICATION:1] [665] [pm-process] Process Alive: coreapplication_obiccs1~OracleBIClusterControllerComponent~BIClusterController~1 (1200161240:4408)

[2012-09-28T13:22:22+08:00] [opmn] [ERROR:1] [] [libopmncustom] Process Ping Failed: 

   特别是 Process Ping Failed ，这一条引人注目，开始怀疑 是进程端口被占用。

  4.检查占用端口。

   ClusterController占用的端口为9706

 查看其端口进程

 netstat  -aon | findstr "9706"

TCP   10.11.1.48:1707  10.11.1.48:9706  SYN_SENT  4348

具体查看进程4348 pid在windows中 服务

 tasklist | findstr "4348"

sawserver.exe      4348 Console  0

这个很正常 sawserver.exe 是opmn整个后台进程的封装。

杀掉sawserver.exe后

。查看opmnctl 的状态

---------------------------------+--------------------+---------+---------
ias-component                    | process-type       |     pid | status
---------------------------------+--------------------+---------+---------
coreapplication_obiccs1          | OracleBIClusterCo~ |    2212 | Stop
coreapplication_obisch1          | OracleBIScheduler~ |    2808 | Alive
coreapplication_obijh1           | OracleBIJavaHostC~ |    3728 | Alive
coreapplication_obips1           | OracleBIPresentat~ |     764 | Alive
coreapplication_obis1            | OracleBIServerCom~ |    3836 | Alive

确定coreapplication_obiccs1 的 pid 为2212 ，该pid 应该封装在sawserver.exe中，但是现在sawserver.exe 已经杀掉了. 如果有与其冲突的进程的话，一定可以找到。

 

D:\0BIEE11G\BIEE\instances\instance1\bin>tasklist|findstr "2212"
wmiprvse.exe                  2212 Console                    0      3,900 K

D:\0BIEE11G\BIEE\instances\instance1\bin>tasklist|findstr "2808"
nqscheduler.exe               2808 Console                    0     57,044 K

D:\0BIEE11G\BIEE\instances\instance1\bin>tasklist|findstr "3728"
java.exe                      3728 Console                    0     29,912 K

D:\0BIEE11G\BIEE\instances\instance1\bin>tasklist|findstr "764"
odbcad32.exe                  5764 RDP-Tcp#4                  1      2,196 K
sawserver.exe                  764 Console                    0    134,356 K

D:\0BIEE11G\BIEE\instances\instance1\bin>tasklist|findstr "3836"
nqsserver.exe                 3836 Console                    0     66,080 K

果然符合我的猜测， pid 2212被 wmiprvse.exe 进程占用了， baidu了一下 该进程，不是OBIEE的系统进程，也不是WindowsOS的核心进程，所以杀掉.

    杀掉后再查看

D:\0BIEE11G\BIEE\instances\instance1\bin>tasklist|findstr "2212"

没有返回结果。

  再次启动OPMN

D:\0BIEE11G\BIEE\instances\instance1\bin>opmnctl startproc ias-component=coreapp
lication_obiccs1
opmnctl startproc: starting opmn managed processes...
================================================================================

opmn id=bi-4lna2lrlna7w:9501
  0 of 0 processes started.
  Processes are already started: instance1~coreapplication_obiccs1~OracleBIClust
erControllerComponent~BIClusterController

D:\0BIEE11G\BIEE\instances\instance1\bin>opmnctl status

Processes in Instance: instance1
---------------------------------+--------------------+---------+---------
ias-component                    | process-type       |     pid | status
---------------------------------+--------------------+---------+---------
coreapplication_obiccs1          | OracleBIClusterCo~ |    4408 | Alive
coreapplication_obisch1          | OracleBIScheduler~ |    2808 | Alive
coreapplication_obijh1           | OracleBIJavaHostC~ |    3728 | Alive
coreapplication_obips1           | OracleBIPresentat~ |     764 | Alive
coreapplication_obis1            | OracleBIServerCom~ |    3836 | Alive

谢天谢地！ 一切正常。

 

 

 

 

 

 

 

 

 

转载于:https://www.cnblogs.com/jerryxing/archive/2012/09/29/2708214.html