Websphere 7.0 was service unable start

故事发生的场景：

    Websphere 7.0 on windows 2008 64bit

故事小人物：wonderful 和徐同学（比我大...）

故事背景：

    用户准备对某个应用作dr 测试，在安装完was 7并重启服务器之后，启动service，但was 的console 无法打开，十几秒后，windows service down。

windows eventlog 中提示：

    Could not determine the process id of the java process. Changing the IBMWAS70Service - SHADCTWAPP90Node02 service status to the "stopped" state. To prevent this error, try recreating this service with the -logRoot parameter.  The value of the logRoot parameter should be the directory in which the server's .pidfile is created.

故事配角（wonderful对was 的认知还停留在小学生认识26个字母的阶段）。首先想到的就是ibm官方或最不济也从谷歌上找找相应的办法。

18摸的其中一个case中，提到相同的场景，具体如下：

Error 134218764 occurs when starting WebSphere services for Rational products after crash or reboot

Problem(Abstract)The IBM WebSphere Application Server services for Rational RequisitePro, Rational ClearCase, and Rational ClearQuest fail to start with an error following a server crash or reboot.SymptomThis problem affects these Rational services and their applicable WebSphere Application Server profiles:    ClearCase Remote Client (CCRC) and ClearQuest Web    IBM WebSphere Application Server V6.1 - cmprofile    ClearQuest Full-Text Search    IBM WebSphere Application Server V6.1 - cqsearchprofile    Report Server for ClearQuest    IBM WebSphere Application Server V6.1 - reportalprofile    RequisitePro RequisiteWeb   IBM WebSphere Application Server V6.1 - RWP ReqWeb ServletYou can experience this behavior with these services after a crash or reboot:    Services are not running as expected after the reboot of the server.    Services will not start when manually requested. An error like this will occur:    Windows could not start the 'IBM WebSphere Application Server V6.1 - RWP ReqWeb Servlet' on Local Computer. For more information, review the System Event Log. If this is a non-Microsoft service, contact the service vendor, and refer to service specific error code 134218764.    The affected WebSphere Application Server profile might not indicate any timestamp changes or log updates that are normally expected during a startup failure. For example, the startServer.log file within this directory of the profile:    <install directory>\IBM\RationalSDLC\common\CM\profiles\<profile name>\logs\server1    ... does not show an update indicating the Windows service associated with the profile tried to start.CauseThis occurs if a stale pid file is left behind in the logs directory. This can cause a conflict with the service startup process.EnvironmentThis problem is observed on Microsoft Windows platforms.Diagnosing the problemAttempt to start the service. Verify if the startServer.log file gets updated. If it is not updated, proceed to the resolution below.Resolving the problem    Locate the server1.pid file in the this directory for the affected profile:    <install directory>\IBM\RationalSDLC\common\CM\profiles\<profile name>\logs\server1    Remove the server1.pid file. Alternatively, you can rename the file to server1.pid.old for backup purposes.    Attempt to start the service again. Upon successful startup, the server1.pid file should be regenerated.

       各位看官姑且先不说 这是was 6.1的问题，事实上这个需要dr的应用也有安装 was6.1，出现过同样的问题，当时我第一反应把was6.1重装,最后重装后也解决问题了，但是还有was7.0的2个服务，居然也有同样的问题，就纳闷了，没人动过环境为啥 3个服务全crash了？

      所以wonderful 决定排查下（其实就是拖上徐同学帮我排查下- -！），但是基于本人was 功力实在难登大雅之堂，我找到了我们was界的武林高手，徐同学，徐同学上来第一个让先检查was环境和系统的java环境，应为was7.0需要jdk 1.6 ，于是乎我在用java -version对was 7.0做检查的时候发现：

这。。。。。这。。。。。 第一反应先去prod环境对比

wonderful心底有种XXniang的冲动（请饶恕我对自己低素质的无知）....徐同学建议我把was 7.0的jdk想各种办法“重装”，介于这是dr 的测试，

首先要体现非全新安装应用的理念，wonderful 首先检查was 7.0下目录的完整性，结果....少了N多的ddl，再次滚粗....

最后，wonderful选择了copy，然后开工,此处省略1万字....copy后，java -version 正常。

接下去的活就..反正就那..那....那...啥....恢复了.....还是省去1万字.....

以上故事纯属真实，如有雷同，纯属你倒霉.....

故事结局： windows service 启动正常，console 访问正常.             

                                                                                                                                                   ----全剧终-----