07.02.08
再破记录
记录,没有最高,只有更高.
一个晚上,迫于系统压力,重起一个系统的数据库,重起后,立刻碰到bug;
没有办法,再来一次.
相对于我们重起一次需要申请15天以上,我估计今天晚上我的操作已经是我们这个team中的记录了.
再破记录,该高兴还是该悲伤,只是希望这样的记录不要再有了,否则我要疯的.
06.23.08
罗列 Oracle 10G OCM 考点
ORACLE 10G OCM考点，罗列之，记录之。
Server Configuration
[ ] Create the database
[ ] Determine and set sizing parameters for database structures
[ ] Create and manage temporary, permanent, and undo tablespaces
[ ] Stripe data files across multiple physical devices and locations
[ ] Configure the database environment to support optimal data access performance
[ ] Create and manage database configuration files
[ ] Create and manage bigfile tablespaces
[ ] Create and manage multiple network configuration files
[ ] Create and configure a listener
[ ] Configure the database instance to support shared server connections
[ ] Set up network tracing
[ ] Manage Oracle network processes (the Listener Control utility)
[ ] Configure the network environment to allow connections to multiple databases
[ ] Use configurationless connections
Enterprise Manager Grid Control
[ ] Install the Enterprise Manager Grid Control software
[ ] Configure the Enterprise Manager repository
[ ] Create Enterprise Manager Grid Control users
[ ] Use Enterprise Manager to modify a database configuration
[ ] Configure Enterprise Manager to modify a database availability
[ ] Create and manage jobs
[ ] Create and monitor alerts
[ ] Create notifications
[ ] Implement Grid Control and Database Control
[ ] Choose the appropriate tablespace type for the intended use
[ ] Create Scheduler jobs
[ ] Create schedules
[ ] Assign jobs to windows
[ ] Create programs
[ ] Create job classes
[ ] Install the Enterprise Manager Grid Control infrastructure
[ ] Deploy Enterprise Manager Grid Control agents
[ ] Configure Grid Control for business requirements
Managing Database Availability
[ ] Create a recovery catalog database
[ ] Configure Recovery Manager
[ ] Use Recovery Manager to perform database backups
[ ] Use Recover Manager to perform complete database restore and recovery operations
[ ] Set Flashback Database parameters
[ ] Monitor Flashback Database logs and statistics
[ ] Perform a Flashback Database operation
[ ] Configure a flash recovery area
Data Management
[ ] Implement fine-grained auditing
[ ] Create a secure application role
[ ] Grant specific privileges for a Flashback Query
[ ] Set parameters for retaining undo
[ ] Implement fine-grained access control
[ ] Create and manage contexts
[ ] Use SQL*Loader
[ ] Implement transportable tablespaces between homogeneous and heterogeneous systems (using different methods of moving files)
[ ] Choose the appropriate partition method (range, hash, list, and composite)
[ ] Choose the appropriate partition key
[ ] Choose appropriate indexing methods (local, global, prefixed, and nonprefixed)
[ ] Perform partition maintenance operations
[ ] Maintain indexes on a partitioned table
[ ] Create and manage LOB segments
[ ] Apply parallelism appropriately
[ ] Tune memory for parallel operations
[ ] Set parallel parameters
Data Warehouse Management
[ ] Create external tables (Data Loader and Data Pump drivers)
[ ] Implement Data Pump export and import jobs for data transfer
[ ] Implement Data Pump to and from remote databases
[ ] Implement Replication/Streams
[ ] Configure and manage master replication
[ ] Configure and manage distributed materialized views
[ ] Configure and manage Streams for replication
Performance Management
[ ] Install Statspack
[ ] Adjust the default interval for statistics collection
[ ] Customize the events that are monitored
[ ] Use ADDM to analyze statistics and to identify performance bottlenecks and fix them
[ ] Implement automatic shared memory management
[ ] Tune memory using manual memory management
[ ] Monitor and tune SGA areas that are not automatically tuned
[ ] Implement automatic SQL execution memory management
[ ] Tune the PGA using work area size parameters
[ ] Use the PGA Advisor
[ ] Create a Database Resource Manager plan with directives (active session count, max execution time)
[ ] Create consumer groups
[ ] Configure consumer group mappings
[ ] Manage Services
[ ] Create and manage objects to accommodate different data access methods (schema tuning)
[ ] Use the SQL Tuning Advisor
[ ] Use the SQL Access Advisor
[ ] Gather Optimizer statistics
[ ] Interpret execution plan
[ ] Use SQL tuning tools and features
Real Application Clusters
[ ] Install Oracle Clusterware
[ ] Install the Oracle Database 10g software
[ ] Implement ASM failure groups
[ ] Create and manage as ASM instance
[ ] Create and manage ASM disk groups
[ ] Configure ASM for the shared disks, and create a clustered database
[ ] Configure archiving
[ ] Configure Services with a primary instance and an available instance
Data Guard
[ ] Use Data Guard Manager and DGMGRL
[ ] Create and utilize a physical standby database
[ ] Create and utilize a logical standby database
[ ] Set up log transport services for various levels or protection
[ ] Configure the network environment to allow communication between the standby database and the primary database
[ ] Open the physical standby database in a “ready-only” state
[ ] Perform a switchover operation and a failover operation
[ ] Implement Data Guard
06.18.08
正式放弃在小黑上安装10g RAC的计划
正式放弃在我的小黑上安装10g RAC的打算。
本来给小黑配置了250g的硬盘，应该是够快了，可惜，小黑的心才1.66，而且内存似乎也不够，经过无数次努力，只要把两个虚拟机打开，cpu温度就直接上85摄氏度左右，而且硬盘转速也抗不住。因此，无数次虚拟机死机后，正式放弃在小黑上搭建linux＋oracle 10g rac的计划。
准备用公司在广州分公司的公网地址来作这个环境，目前已知的该环境是2颗双核的cpu，内存8g，硬盘我占了60g，似乎是scsi盘，硬盘速度足够，将转战到那里去整，反正远程过去也蛮方便的，远程桌面而已。公司能提供如此环境，真的是很开心。
这次总应该能顺利搭建起我的rac环境吧。不过无数次的折腾后，发现该碰到的问题都碰到过，实在是太有意思了。
碰到问题，解决之，何尝不是一种快感！
06.14.08
作技术，什么都可能发生
最近的怪事蛮多。
首先是那天我晚班，下午去公司，同事说下午2点左右地震了，我靠，那可是6月12日，难道是汶川地震1个月的纪念，再问下去，原来是那个时候一台伟大的DS8300可能是站着太累了，居然在机房去睡了会，1.4吨的家伙，这么一睡，那动作和地震差不多。晚饭后去机房看望了下这哥们，嗯，这是第一次我有意识的去看8300，结果还是刚睡醒的，这种看8300倒地的机会还真不多，可惜我去看望的时候8300已经站起来了，不过没人敢给它上电了。有多少人见过8300倒地呀？
今天晚上在家值班的时候，收到电话，晚上要重新起停一个rac库。不是很重要的系统，所以服务器也不是很好，IBM P560Q而已，也没怎么在意，这种活已经习惯了，都不会紧张了，先把库拉掉，很正常，然后IBM的哥们在机房给加内存，装好了，HA起来，正常，然后我拉库。拉呀，第一个节点起来正常，拉第二个节点呗，启动到nomount就不动了，检查alert.log
at Jun 14 02:46:52 2008
lmon registered with NM - instance id 2 (internal mem no 1)
Sat Jun 14 02:52:41 2008
Reconfiguration started (old inc 0, new inc 1)
List of nodes:
1
Global Resource Directory frozen
one node partition
Communication channels reestablished
Master broadcasted resource hash value bitmaps
Non-local Process blocks cleaned out
Resources and enqueues cleaned out
Resources remastered 0
0 GCS shadows traversed, 0 cancelled, 0 closed
0 GCS resources traversed, 0 cancelled
set master node info
Submitted all remote-enqueue requests
Update rdomain variables
Dwn-cvts replayed, VALBLKs dubious
All grantable enqueues granted
0 GCS shadows traversed, 0 replayed, 0 unopened
Submitted all GCS remote-cache requests
0 write requests issued in 0 GCS resources
0 PIs marked suspect, 0 flush PI msgs
也没啥好看的，把先拉起来的节点down掉，2号节点就立刻起来了。这问题好像以前见过，我记得好像是如果参数cluster_interconnects不对的话，会有这种问题，嗯，那就整呗，这个参数也正常，每个instance单独起来都会对。难道是，莫非是见鬼了，ok，再想想，似乎两个主机交换的hacmp的网络会引起这种问题，先检查tty0，正常，再检查网卡的参数MTU，也正常。
无论如何，我一定坚信，这个是HA的问题或者是网络的问题，明显资源有问题。再看看metalink，发现也是这些东西，但是metalink说的这些都没有问题的，继续折磨IBM的哥们，让他们去查网络，终于，10分钟后，有好消息了，RAC的网络光纤出现了问题，我靠，不就是下个电吗，有这么麻烦的加内存动的是前面，又不是后面，光纤，你咋坏了列。赶紧的兄弟们换光纤线，重新拉库。正常了。
不是8300倒地，就是光纤不行，是我该去烧香，还是IBM的哥们该去了列？