The words are in Chinese. For sure, the script is in English.
Here is the link to the blog.
Thursday, August 09, 2012
Sunday, February 05, 2012
How to install Grid Control 11gR1 on Linux 5 - complete reference
Oracle Enterprise
Manager Grid Control (EMGC) is
a system management software that delivers centralized monitoring,
administration, and life cycle management functionality for the complete IT
infrastructure, including systems running Oracle and non-Oracle technologies,
from a single console.
Repository database
version needs to be any one of the following versions:
10.2.0.4.0
11.1.0.7.0
11.2.0.1.0
11.2.0.2.0
Oracle official documents about EMGC 11g:
10.2.0.4.0
11.1.0.7.0
11.2.0.1.0
11.2.0.2.0
Oracle official documents about EMGC 11g:
Oracle® Enterprise Manager Grid
Control Basic Installation Guide 11g Release 1 (11.1.0.1.0) (E15838)
|
Oracle Useful Metalink Notes about EMGC 11g:
Master Note for Grid
Control 11.1.0.1.0 Installation and Upgrade [ID 1067438.1]
Complete Steps to
Install 11.1.0.1 OMS Using 11.2 Database on 64 Bit RedHat 5 or OEL5 [ID
1130958.1]
Some useful sites about the EMGC 11g installation:
- How to Build a Grid Control Environment (11g) (Step by step documentation from oracle-wiki, it is based on Oracle Linux 5)
- OEM Grid Control 11g(11.1.0.1.0) Installation (A popular Step by step documentation from learnwithme11g - SRIKRISHNA MURTHY, very good note about the installation)
- OEM Grid Control 11gR1
Installation on Oracle Linux 5.5 (64-bit) (Step by step
documentation from oracle-base, based on
Oracle Linux 5.5, including OS, Java, Database, Weblogic and GC
installations)
- Installation Guide Oracle
Enterprise Manager Grid Control 11g on RHEL 5.5 x86-32 (a
PDF document about the installation including
the screenshot, from Martin Decker, an independent Oracle
consultant. )
- Installing Oracle Enterprise
Manager Grid Control 11g Release 1 (11.1.0.1.0) Linux x86 (from Oracle APPS DBA, based on Oracle
Linux 5.4)
- Oracle Grid Control 11gR1
Installation (Step by step documentation from Justin)
- Oracle Enterprise Manager Grid
Control Install 11g R1 11.1.0.1 (a YouTube video about GC
11g Installation from MASTERSCHEMA)
- Oracle DBA – Tips and Techniques (em-grid-control) (a blog site about Oracle stuff from Gavin Soorma, an Australian)
Oracle official site about EMGC (now focus on 12c):
Wednesday, November 09, 2011
New ORA-4031 Troubleshooting Tool
Yesterday (Nov.09, 2011) Oracle delivered a web seminar to talk about the new tool. The tool has similar purpose of "ORA-600/ORA-7445 Error Look-up Tool" and can provide DBA an easy way to diagnose the ORA-4031 errors.
By uploading your problematic database's alert log file and trace files, the tool can automatically provide the issue identification and suggested solution based on its established guidelines. Actually, you can get to read those guidelines without really uploading anything. So, you can even use the guideline to troubleshoot the database by yourself.
Here is the example information taken from the guideline:
Also, in the seminar, the presenter mentioned Oracle is also building the ORA-4300 troubleshooting tool. By promoting these tools, Oracle is not only helping DBA to do the work by themselves, but also saving money from support. In the end, Oracle can make more profit and get more customers.
By uploading your problematic database's alert log file and trace files, the tool can automatically provide the issue identification and suggested solution based on its established guidelines. Actually, you can get to read those guidelines without really uploading anything. So, you can even use the guideline to troubleshoot the database by yourself.
Here is the example information taken from the guideline:
Fault: Insufficient
SGA Free Memory at Startup
|
|
Your instance SGA configuration has been flagged as having too little
memory available for the automatic memory manager to make adjustments in the
SGA component sizes.
This issue could occur if in the init.ora parameters of your Alert log, (shared_pool_size + large_pool_size + java_pool_size + db_keep_cache_size + streams_pool_size + db_cache_size) / sga_target is greater than 90%. With the introduction of automatic memory management (ASMM in 10g, AMM in 11g), i.e. when SGA_TARGET > 0 or MEMORY_TARGET > 0, memory granules can be moved from one SGA component to another, depending on which component has an immediate need to grow. The memory components are: shared pool, buffer cache, streams pool, large pool, java pool, log buffer. The most common sharing of memory granules occur between the shared pool and the buffer cache and so these components can increase and decrease in size. In order to limit how much any component will decrease in size, we set a minimum value by specifying a size in the parameter file. For example, when SGA_TARGET >0 or MEMORY_TARGET > 0, then SHARED_POOL_SIZE=nnnn and B_CACHE_SIZE=nnnn become minimum sizes of these components. The sum of all the minimum sizes determine the minimum size of the SGA at start up. If the SGA_TARGET (or MEMORY_TARGET) is only slightly larger than the sum of the minimum settings, there will be very little memory to do any resizing, and this can lead to a ORA-4031. |
|
Relevant Observation(s):
|
|
Resolution Action Plan
|
|
1) Reduce the minimum values for the dynamic SGA components to allow
memory manager to make changes as needed
For example, either set: shared_pool_size=new value in the database INIT.ORA file, or connect as SYSDBA and issue statements: alter system set shared_pool_size=new value scope=both; NOTE: Explicit settings (as minimums) can be changed while the database is online. and 2) Modify auto-tuned configuration 10g: Increase value of SGA_TARGET to provide more SGA memory Either set: sga_target=new value in the database INIT.ORA file, or connect as SYSDBA and issue statements: alter system set sga_target=new value scope=both; Note: It will be necessary to restart the instance. That will mean that associated database will have to schedule an outage as well. 11g: Increase value of MEMORY_TARGET if used. If SGA_MAX_SIZE has been set remove the explicit setting and allow MEMORY_TARGET to manage the SGA sizing. Either set: memory_target=new value (remove explicit setting if applicable) sga_max_size=value in the database INIT.ORA file, or connect as SYSDBA and issue statements: alter system set memory_target=new value scope=both; (if explicitly setting SGA_MAX_SIZE) alter system reset sga_max_size scope=spfile; Note: It will be necessary to restart the instance. That will mean that associated database will have to schedule an outage as well. or 3) both of the above. 4) See these notes for further information regarding automatic memory configuration. NOTE:781630.1 - How to configure ASMM (Auto Memory) on 10g and 11g - Best Practices NOTE:452512.1- 11g MEMORY_TARGET Parameter Dependency NOTE:443746.1- Automatic Memory Management(AMM) on 11g NOTE:295626.1- How To Use Automatic Shared Memory Management (ASMM) In Oracle10g |
Also, in the seminar, the presenter mentioned Oracle is also building the ORA-4300 troubleshooting tool. By promoting these tools, Oracle is not only helping DBA to do the work by themselves, but also saving money from support. In the end, Oracle can make more profit and get more customers.
Saturday, November 05, 2011
Oracle Internal Research深入研究Oracle内部原理
以下是刘相兵(Maclean Liu)个人原创或总结的一些深入研究Oracle Internal内部原理的文章post列表:
Oracle Internal Event:10200 Consistent Read诊断事件
了解Oracle RAC Brain Split Resolution
了解Oracle内核代码层的作用
了解你所不知道的SMON功能(九):维护MON_MODS$字典基表
了解你所不知道的SMON功能(八):Transaction Recover
了解你所不知道的SMON功能(七):清理IND$字典基表
了解你所不知道的SMON功能(六):Instance Recovery
了解你所不知道的SMON功能(五):Recover Dead transaction
了解你所不知道的SMON功能(四):维护col_usage$字典基表
了解你所不知道的SMON功能(三):清理obj$基表
了解你所不知道的SMON功能(二):合并空闲区间
了解你所不知道的SMON功能(一):清理临时段
Know more about commit
Find INTCOL#=1001 in col_usage$?
Buffer Lock Mode and Compatibilities
关于V$OPEN_CURSOR
Advanced Diagnostic using oradebug dumpvar
还原真实的cache recovery
Oracle内部视图:x$targetrba
Oracle内部视图:x$ktfbfe
Oracle内部视图:x$ktfbue
Only ARCH Bgprocess may create archivelog?
Oracle Latch:一段描绘Latch运作的伪代码
Who pulls the background process dbwr’s trigger?
基于行跟踪的ROWDEPENDENCIES ORA_ROWSCN信息
Oracle闩:Cache Buffers chains
Latches and Tuning:Latches
Latches and Tuning:Redo Log Buffer and Latches
Latches and Tuning:Buffer Cache
Latches and Tuning:The Library Cache
Oracle Database Links Master Class
Oracle等待事件:Data file init write
What’s the Point of Oracle Checkpoints?
How does RECORDLENGTH affect your exp speed?
Oracle网络TNS协议介绍(revised)
logfile switch causes incremental checkpoint?
Know more about redo log buffer and latches
How does cpu_count parameter affect instance?
How to trigger ORA-00600,ORA-7445 by manual
Basics of C code within the Oracle kernel
Know more about Oracle Latches
Know more about Buffer Cache and Latch
Know More About Libarary Cache and Latches
Oracle内部视图:X$BH
Know more about DML
Known Oracle Internal Stack Call Meaning
深入了解ASMM
_shared_pool_reserved_pct or shared_pool_reserved_size with ASMM
PL/SQL Virtual Machine Memory Usage
Slide:深入了解Oracle自动内存管理ASMM by Maclean Liu
Know GCS AND GES structure size in shared pool
How does SGA/PGA allocate on AMM?
理解Oracle在AIX平台上的内存使用
Probe how does your PGA consume
ora-4031 and “obj stat memory” component in Shared Pool
直接路径读取对于延迟块清除的影响
How to make BBED(Oracle Block Brower and EDitor Tool) on Unix/Linux/Windows
隐藏参数_high_priority_processes与oradism
Oracle等待事件Enqueue CI:Cross Instance Call Invocation
Oracle常用的几个父栓
ddl操作是否会产生undo?
kgl simulator,共享池simulator以及缓冲simulator相关解释
Saturday, October 29, 2011
OOW 2011
If you didn't attend this year's OOW, you still can download the presentation for free. Click here.
Tuesday, September 13, 2011
Demo about Oracle Parsing
This is a good demo from Andrew Holdsworth, Senior Director of Oracle’s Real World Performance (RWP) team. He showed us how soft parsing, hard parsing and direct logon/off from app (no using connection pool) to degrade the db performance (comparing with that without any parsing at all).
Sunday, May 01, 2011
Monday, January 05, 2009
Oracle ORION
ORION is a standalone tool for calibrating the I/O performance for storage systems that are intended to be used for Oracle databases.
To find out the detail and play with it, you can find the detail here.
To find out the detail and play with it, you can find the detail here.
Friday, January 02, 2009
DBA 2.0
While the whole world is talking about the web 2.0, DBA 2.0 is also not a NEW thing. To learn more about it, here is a good article: Performing an Oracle DBA 1.0 to DBA 2.0 Upgrade. Actually, it was a dialogue between 2 Oracle gurus.
Tuesday, December 23, 2008
pga_aggregate_target and db_cache_size
In his blog article Analysing Statspack(8), Jonathan Lewis has talked about 2 rules:
pga_aggregate_target - A comfortable figure for an OLTP system would be in the 2M - 4M range per process.
db_cache_size - is always at least as large as 1M per process.
Take a note first and check reason later.
pga_aggregate_target - A comfortable figure for an OLTP system would be in the 2M - 4M range per process.
db_cache_size - is always at least as large as 1M per process.
Take a note first and check reason later.
Wednesday, November 26, 2008
Oracle 9.2.0.8 and 10.2.0.4 Patch download list
Oracle 9.2.0.8 patch:
ftp://updates.oracle.com/4547809/p4547809_92080_AIX64-5L.zip
ftp://updates.oracle.com/4547809/p4547809_92080_AXP.zip
ftp://updates.oracle.com/4547809/p4547809_92080_HP64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_HPUX-IA64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_LINUX-S390.zip
ftp://updates.oracle.com/4547809/p4547809_92080_LINUX.zip
ftp://updates.oracle.com/4547809/p4547809_92080_Linux-IA64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_Linux-x86-64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_MVS.zip
ftp://updates.oracle.com/4547809/p4547809_92080_SOLARIS.zip
ftp://updates.oracle.com/4547809/p4547809_92080_SOLARIS64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_TRU64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_WINNT.zip
ftp://updates.oracle.com/4547809/p4547809_92080_WINNT64.zip
Oracle 10.2.0.4 patch:
ftp://updates.oracle.com/6810189/p6810189_10204_AIX5L.zip
ftp://updates.oracle.com/6810189/p6810189_10204_HPUX-IA64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_HPUX-64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Solaris-64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Linux-x86-64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Linux-x86.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Win32.zip
ftp://updates.oracle.com/6810189/p6810189_10204_MSWIN-x86-64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_AIX64-5L.zip
ftp://updates.oracle.com/4547809/p4547809_92080_AXP.zip
ftp://updates.oracle.com/4547809/p4547809_92080_HP64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_HPUX-IA64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_LINUX-S390.zip
ftp://updates.oracle.com/4547809/p4547809_92080_LINUX.zip
ftp://updates.oracle.com/4547809/p4547809_92080_Linux-IA64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_Linux-x86-64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_MVS.zip
ftp://updates.oracle.com/4547809/p4547809_92080_SOLARIS.zip
ftp://updates.oracle.com/4547809/p4547809_92080_SOLARIS64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_TRU64.zip
ftp://updates.oracle.com/4547809/p4547809_92080_WINNT.zip
ftp://updates.oracle.com/4547809/p4547809_92080_WINNT64.zip
Oracle 10.2.0.4 patch:
ftp://updates.oracle.com/6810189/p6810189_10204_AIX5L.zip
ftp://updates.oracle.com/6810189/p6810189_10204_HPUX-IA64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_HPUX-64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Solaris-64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Linux-x86-64.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Linux-x86.zip
ftp://updates.oracle.com/6810189/p6810189_10204_Win32.zip
ftp://updates.oracle.com/6810189/p6810189_10204_MSWIN-x86-64.zip
Thursday, November 20, 2008
Search Oracle Book on Google Book and Read online
Nowadays there is no much time to read Oracle books (I mean the real books). We rather to search in Metalink, Google or other Oracle related sites. As we all know, those books are not cheap. If we do want to buy them, we would like the company to pay for the bills:-)
Actually, we can find lots of Oracle books for online reading. Here is an example. You can search Oracle book by using Google Book:
http://books.google.com/books?q=Oracle&lr=&sa=N&start=0
You can even fine-tune your search. Here are some books for your reference:
Oracle Essentials: Oracle Database 10g
Oracle Performance Tuning
Oracle High Performance Tuning for 9i and 10g
Oracle Performance Troubleshooting: With Dictionary Internals SQL & Tuning
Be aware: you can NOT read full pages of the above books. If you feel those books really good, you will have to buy by yourself or company. Anyway, enjoy!
Actually, we can find lots of Oracle books for online reading. Here is an example. You can search Oracle book by using Google Book:
http://books.google.com/books?q=Oracle&lr=&sa=N&start=0
You can even fine-tune your search. Here are some books for your reference:
Oracle Essentials: Oracle Database 10g
Oracle Performance Tuning
Oracle High Performance Tuning for 9i and 10g
Oracle Performance Troubleshooting: With Dictionary Internals SQL & Tuning
Be aware: you can NOT read full pages of the above books. If you feel those books really good, you will have to buy by yourself or company. Anyway, enjoy!
Thursday, April 27, 2006
How to Misuse SQL's FROM Clause
http://www.onlamp.com/lpt/a/5220
Stéphane Faroult first discovered the relational model and the SQL language in 1983.
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnsqlpro05/html/sp05a1.asp
http://searchoracle.techtarget.com/ateQuestionNResponse/0,289625,sid41_cid611283_tax301455,00.html
Stéphane Faroult first discovered the relational model and the SQL language in 1983.
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnsqlpro05/html/sp05a1.asp
http://searchoracle.techtarget.com/ateQuestionNResponse/0,289625,sid41_cid611283_tax301455,00.html
Wednesday, February 01, 2006
Default Kernel Parameters for Sun OS 9.X Database
Metalink Note:187273.1 Type: PROBLEM
Last Revision Date: 21-OCT-2005 Status: PUBLISHED
Solution Description:
---------------------
This article is to define the default kernel parameter settings for the
Sun OS/Solaris Operating system running Oracle 8.X Enterprise Edition.
Solution Explanation:
---------------------
In Sun OS the /etc/system file is the kernel parameter file.
The current kernel parameters values can be displayed by using the
'sysdef -i' command.
- Refer to Note 70688.1"Location and Dispay of kernel parameters" for more
information on uderstanding the 'sysdef -i'output.
======================================================================
UNIX Kernel Parameters:
-----------------------
SEMMNI 100
- Defines the maximum number of semaphore sets in the entire system.
SEMMNS 256
- Defines the maximum semaphores on the system. This setting is a
minimum recommended value, for initialinstallation only. The SEMMNS
parameter should be set to the sum of the PROCESSES parameter for each Oracle
database, adding the largest one twice, and then adding an additional 10 for
each database.
SEMMSL 256
- Defines the minimum recommended value, for initial installation only.
SHMMAX 4294967295
- Defines the maximum allowable size of one shared memory segment.
(4 GB = 4294967295)
SHMMIN 1
- Defines the minimum allowable size of a single shared memory segment.
SHMMNI 100
- Defines the maximum number of shared memory segments in the entire system.
SHMSEG 10
- Defines the maximum number of shared memory segments one process can attach.
Note: These are minimum kernel requirements for Oracle9i. If you have previously
tuned your kernel parameters to levels equal to or higher than these values,
continue to use the higher values. A system restart is necessary for kernel
changes to take effect.
------------------------------------------------------------------------------
* Example of Oracle required parameters:
set shmsys: shminfo_shmseg = 10
set shmsys: shminfo_shmmin = 1
set shmsys: shminfo_shmmni = 100
set semsys: seminfo_semmns = 256
set semsys: seminfo_semmni = 4096
set semsys: seminfo_semmsl = 256
set shmsys: shminfo_shmmax = 4294967295
Last Revision Date: 21-OCT-2005 Status: PUBLISHED
Solution Description:
---------------------
This article is to define the default kernel parameter settings for the
Sun OS/Solaris Operating system running Oracle 8.X Enterprise Edition.
Solution Explanation:
---------------------
In Sun OS the /etc/system file is the kernel parameter file.
The current kernel parameters values can be displayed by using the
'sysdef -i' command.
- Refer to Note 70688.1"Location and Dispay of kernel parameters" for more
information on uderstanding the 'sysdef -i'output.
======================================================================
UNIX Kernel Parameters:
-----------------------
SEMMNI 100
- Defines the maximum number of semaphore sets in the entire system.
SEMMNS 256
- Defines the maximum semaphores on the system. This setting is a
minimum recommended value, for initialinstallation only. The SEMMNS
parameter should be set to the sum of the PROCESSES parameter for each Oracle
database, adding the largest one twice, and then adding an additional 10 for
each database.
SEMMSL 256
- Defines the minimum recommended value, for initial installation only.
SHMMAX 4294967295
- Defines the maximum allowable size of one shared memory segment.
(4 GB = 4294967295)
SHMMIN 1
- Defines the minimum allowable size of a single shared memory segment.
SHMMNI 100
- Defines the maximum number of shared memory segments in the entire system.
SHMSEG 10
- Defines the maximum number of shared memory segments one process can attach.
Note: These are minimum kernel requirements for Oracle9i. If you have previously
tuned your kernel parameters to levels equal to or higher than these values,
continue to use the higher values. A system restart is necessary for kernel
changes to take effect.
------------------------------------------------------------------------------
* Example of Oracle required parameters:
set shmsys: shminfo_shmseg = 10
set shmsys: shminfo_shmmin = 1
set shmsys: shminfo_shmmni = 100
set semsys: seminfo_semmns = 256
set semsys: seminfo_semmni = 4096
set semsys: seminfo_semmsl = 256
set shmsys: shminfo_shmmax = 4294967295
9.2.0.4 databases to 9.2.0.6, 9.2.0.7 or wait for testing?
http://searchoracle.techtarget.com/expert/KnowledgebaseAnswer/0,289625,sid41_gci1157663,00.html?track=sy40
My name is Steve and I'm a Lead Oracle DBA. My question revolves around the quarterly security patches and the best 9i release of Oracle to be on. I have over 300 9.2.0.4 databases. We just started creating 9.2.0.6 databases with an eye toward going to 10.2 in the first quarter next year. My question/issue is this: Oracle released 9.2.0.7 recently, and my concern is that they will stop patching 9.2.0.6 in the near future (not sure of that date). I'd like to be on the terminal release of 9i but that's been a moving target. Anyway, should I upgrade my existing 9.2.0.4 databases to 9.2.0.6, 9.2.0.7 or wait for testing, etc. of 10.2 and just move them all to that release? Any advice would be helpful. Thanks.
This question posed on 10 January 2006
>This is a subject I have quite an interest in so I could probably spend hours discussing it. This patching issue is still relatively new to most DBAs and it can be especially painful if there are a large number of databases to support, as is the case here. I work for an IT consulting company, but I just spent the last two years at a large client and we had to tackle this exact problem. This client had more than 200 Oracle databases that had to be patched on a regular basis. With Sarbanes-Oxley (and other) regulatory compliance legislation, patching databases is no longer an option, but a necessity.
Your concern that patches will no longer be available for Oracle 9.2.0.6 is quite valid -- this, in fact, will happen one day soon. So, in my opinion, you need to develop a strategy that balances the practicality of patching but can tolerate some risk. For example, you can apply the appropriate CPU patches once per year (say, in late spring) and then plan to upgrade databases to the next release in the fall. In theory, the latest release will contain the latest CPUs. It's nearly impossible to upgrade 300 databases more than once per year; as well, it would be impossible to apply all four CPUs to these databases considering you would require outages which can be very difficult to obtain on production systems.
Whatever strategy you choose, make sure that it works for your organization and that you can justify it. Also, document the strategy and your rationale for choosing it.
My name is Steve and I'm a Lead Oracle DBA. My question revolves around the quarterly security patches and the best 9i release of Oracle to be on. I have over 300 9.2.0.4 databases. We just started creating 9.2.0.6 databases with an eye toward going to 10.2 in the first quarter next year. My question/issue is this: Oracle released 9.2.0.7 recently, and my concern is that they will stop patching 9.2.0.6 in the near future (not sure of that date). I'd like to be on the terminal release of 9i but that's been a moving target. Anyway, should I upgrade my existing 9.2.0.4 databases to 9.2.0.6, 9.2.0.7 or wait for testing, etc. of 10.2 and just move them all to that release? Any advice would be helpful. Thanks.
This question posed on 10 January 2006
>This is a subject I have quite an interest in so I could probably spend hours discussing it. This patching issue is still relatively new to most DBAs and it can be especially painful if there are a large number of databases to support, as is the case here. I work for an IT consulting company, but I just spent the last two years at a large client and we had to tackle this exact problem. This client had more than 200 Oracle databases that had to be patched on a regular basis. With Sarbanes-Oxley (and other) regulatory compliance legislation, patching databases is no longer an option, but a necessity.
Your concern that patches will no longer be available for Oracle 9.2.0.6 is quite valid -- this, in fact, will happen one day soon. So, in my opinion, you need to develop a strategy that balances the practicality of patching but can tolerate some risk. For example, you can apply the appropriate CPU patches once per year (say, in late spring) and then plan to upgrade databases to the next release in the fall. In theory, the latest release will contain the latest CPUs. It's nearly impossible to upgrade 300 databases more than once per year; as well, it would be impossible to apply all four CPUs to these databases considering you would require outages which can be very difficult to obtain on production systems.
Whatever strategy you choose, make sure that it works for your organization and that you can justify it. Also, document the strategy and your rationale for choosing it.
Monday, January 16, 2006
How to find obsolete parameters in Oracle
By querying V$OBSOLETE_PARAMETER.
If the value in column ISSPECIFIED is 'TRUE' then it is specified in the init.ora file.
select * from V$OBSOLETE_PARAMETER
where ISSPECIFIED = 'TRUE'
/
If the value in column ISSPECIFIED is 'TRUE' then it is specified in the init.ora file.
select * from V$OBSOLETE_PARAMETER
where ISSPECIFIED = 'TRUE'
/
Monday, January 09, 2006
Thursday, December 22, 2005
Salesforce "failover"?
CNET reported on December 21, 2005: A Salesforce.com outage lasting nearly a day cut off access to critical business data for many of the company's customers on Tuesday in what appears to be Salesforce's most severe service disruption to date.
Salesforce, which has been growing rapidly, has undertaken efforts to bolster its computing infrastructure. For instance, it has configured its database to run on four different computers so if a machine fails, others will pick up the slack, Francis said. But the "failover" feature didn't prevent Tuesday's problems.
Salesforce's database supplier helped to restore service, Francis said. While he declined to identify who that supplier was, he did identify Oracle as Salesforce's biggest database supplier.
To see the detail, go to http://news.com.com/Salesforce+outage+angers+customers/2100-1012_3-6004625.html?tag=nefd.top
Salesforce, which has been growing rapidly, has undertaken efforts to bolster its computing infrastructure. For instance, it has configured its database to run on four different computers so if a machine fails, others will pick up the slack, Francis said. But the "failover" feature didn't prevent Tuesday's problems.
Salesforce's database supplier helped to restore service, Francis said. While he declined to identify who that supplier was, he did identify Oracle as Salesforce's biggest database supplier.
To see the detail, go to http://news.com.com/Salesforce+outage+angers+customers/2100-1012_3-6004625.html?tag=nefd.top
Tuesday, December 06, 2005
Manually Resolving In-Doubt Transactions: Different Scenarios
NOTE1: If using Oracle 9i and DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY fails with
ORA-30019: Illegal rollback Segment operation in Automatic Undo mode, use the following workaround
SQL> alter session set "_smu_debug_mode" = 4;
SQL>execute DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY('local_tran_id');
select * from dba_2pc_pending
/
SQL> select LOCAL_TRAN_ID, STATE, MIXED, ADVICE from dba_2pc_pending;
LOCAL_TRAN_ID STATE MIX A
---------------------- ---------------- --- -
3.7.99084 prepared no
http://www-rohan.sdsu.edu/doc/oracle/server803/A54647_01/ch4e.htm
COMMIT FORCE '3.7.99084';
SQL> select LOCAL_TRAN_ID, STATE, MIXED, ADVICE from dba_2pc_pending;
LOCAL_TRAN_ID STATE MIX A
---------------------- ---------------- --- -
3.7.99084 forced commit no
SQL> select * from dba_pending_transactions;
FORMATID
----------
GLOBALID
--------------------------------------------------------------------------------
BRANCHID
--------------------------------------------------------------------------------
48801
34A257C2BC134A007FFD
73616D705841436F6E6E506F6F6C
alter session set "_smu_debug_mode" = 4;
execute DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY('3.7.99084');
SQL> select * from dba_pending_transactions;
no rows selected
SQL> select * from dba_pending_transactions;
no rows selected
ORA-30019: Illegal rollback Segment operation in Automatic Undo mode, use the following workaround
SQL> alter session set "_smu_debug_mode" = 4;
SQL>execute DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY('local_tran_id');
select * from dba_2pc_pending
/
SQL> select LOCAL_TRAN_ID, STATE, MIXED, ADVICE from dba_2pc_pending;
LOCAL_TRAN_ID STATE MIX A
---------------------- ---------------- --- -
3.7.99084 prepared no
http://www-rohan.sdsu.edu/doc/oracle/server803/A54647_01/ch4e.htm
COMMIT FORCE '3.7.99084';
SQL> select LOCAL_TRAN_ID, STATE, MIXED, ADVICE from dba_2pc_pending;
LOCAL_TRAN_ID STATE MIX A
---------------------- ---------------- --- -
3.7.99084 forced commit no
SQL> select * from dba_pending_transactions;
FORMATID
----------
GLOBALID
--------------------------------------------------------------------------------
BRANCHID
--------------------------------------------------------------------------------
48801
34A257C2BC134A007FFD
73616D705841436F6E6E506F6F6C
alter session set "_smu_debug_mode" = 4;
execute DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY('3.7.99084');
SQL> select * from dba_pending_transactions;
no rows selected
SQL> select * from dba_pending_transactions;
no rows selected
Subscribe to:
Posts (Atom)