如何通过加载系统参数故意使Oracle VM崩溃

我使用下面的rb.sql脚本。

-- The Following script shows the time it takes to insert 1 million rows in a database. -- I used it to determine that: -- -- 1) Optimized Databases perform in relation to the hardware they run on, -- and all the databases on 1 platform run within 10% of each other. -- 2) When Rollback, Temp or Redo logs share the same underlying disks as -- the "alan" table, the insert takes up to twice as long. -- 3) When the rollback segments, Temp segments or redo log buffers are > 20 meg as -- opposed to less than 1 meg, the inserts run 2 to 5 times faster. -- By making the log_buffer and Rollback extents sizes close to 32k, -- the inserts run up to 9 times slower. set feedback off set termout off drop table alan; set termout on create table alan (name varchar2(30)) storage (initial 32k next 1m) logging noparallel storage(buffer_pool default); insert into alan values ('alan'); insert into alan values ('alan1'); insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; insert into alan select * from alan; set timing on insert /*+ PPEND */ into alan logging select * from alan /* This line shows the elapsed time. */ ; set timing off -- select max(sequence#) from v$loghist; set feedback on set termout off --drop table alan; set termout on 

我用rb16.sql调用它：

 @rb @rb @rb @rb @rb @rb @rb @rb @rb @rb @rb @rb @rb @rb @rb @rb 

我用call_rb16.sql调用rb16.sql：

 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 @rb16 

由于rb.sql脚本插入到alan表中，因此可以创build多个oracle用户并从多个login插入。 rb.sql脚本识别瓶颈。

只需使用Linux 压力实用程序select性地加载部分或全部系统function就非常简单。 它可以通过YUM下载…

把stress想象成一个负载生成器。 防止系统崩溃是一个不同的努力。 你可以用它来看看事情是如何在负载下运行的，但是我怀疑来自压力工具的任何健全的configuration都会导致系统崩溃。

另请参阅： * nix中的最大CPU负载