Shared SQL area: It is used to store the frequently accessed tables, element and SQL
statements
It consists of DD cache and SQL cache
DB Buffer: It is used to store the conetent i.e fetched from the database, it has its own
size
-DB buffer size should large enough to hold the user request data
-Each and every request should get response from DB buffer
Redolog Buffer:
It is used for fast commit processing; it has its own size i.e., 1mb up to oracle 9i, from
oracle 10g it has 4 MB
Fast commit –it ensures that user transaction is updated into database
-user transactions are committed only when they are written to redolog files
-user gets updated the response immediately when it is updated in the redolog groups (A
& B) on the disk
Control files
These are used to update the information dynamically about the database structure
changes
-It consists of the paths, log sequence no, system changes number, and it is essential to
restore and recover the database.
-Control files needs to be placed in 3 different physical locations, which can be extended
up to 6.
-This is to ensure that atleast one control file is available to restore and recover the db
Cntrl<SID>.dbf located in cntrl directory
Datafiles
Datafile are the physical files at the OS Level
Oracle has its own background process
Log writer:
It is used to write the content from redolog buffer (memory) to online redolog buffer
(disk)
It writes when
1/3rd of the log buffer is full
If the user commits size more that 1MB
Before DB writes, writes to the db it triggers
When a user commits a transaction it triggers
Database writer (DBWR):
It is used to write the contents which are marked as dirty buffers to data buffers
It is triggered
For every 15 seconds
When the checkpoint occurred again it writes
When user request (db process) does not find space database buffers to host the fetched
content it trigger
Check point:
It is used to maintain the consistence of the consistency of the system by triggering log
writer and db writer
It is triggered for every 60 seconds which can be increased or decreased
SMON:
It is used to bring the system into a consistency stage when the system is terminated or
uploaded during power failure.
SMON does the following tasks during startup the database
Before the db is opened, SMON reads the online redolog files and redo the changes in the
database as per commit
SMON reads roll back or undo segments to roll back or undo changes
SMON releases the user requests and locks
SMON brings the system in consistent mode
PMON:
It is used to monitor the process and ensure that all the process are released, before
openings the database
Archive Process:
It is used to write the contents from OriLog A & B to oraarch directory
When orilog A is full log switch occur and log files are written to OriLog B and A and
the sequence number is increased
Sap data 1….Sap data 4: These are the physical files
SAPTRACE: It is used to write the log files of all the database operation
SAPCHECK: It is written with log files when a command to check the database is issued
SAPBACKUP: It writes the log files of the backup, when the backup is triggered to tape
This is also a directory for ac backup for the disk backup is scheduled
SAPARCH: It is used to host the log files that are generated during the log backup
SAPREORG: It is used to host the data import and export during re-organization
-Orilog A and Orilog B: These are online redolog files with atleast one member in each
group
SAP recommends having 2 groups that should be placed in 2 different disks
The size of the log files varies from company to company, 25-50 MB
ARCHIVE: on production systems the archive log mode should be set to yes, if it is set to
yes, org process writes from LOG A and LOG B to ora arch directory
But on training and demo systems log mode can be set to no, so that archive process
never gets triggered
When OriLog A is filled it will overwrite the content of OriLog B
During abnormal termination we cannot restore the system to the point of consistency
Oracle startup and shutdown process
When db starts or startdb command is initialized the parameter are read from
init<SID>.ora which resides in d:\oracle\ora92\
It consists of
-SGA (Shared Global Area)
database buffer: db_block_buffer=26080
-Log-buffer
-shared_pool_size
-control file (3 locations are specified, 6 location can be specified)
-db_block_size=8192
-archive destination
There are 3 types of parameters:
Tuning Parameters
-This parameter is used to tune the performance of the system
Changeable Parameters
-These parameters can be changes based on the requirements like process, change the
archive log mode, control file location etc
Static Parameters
-Most of the parameters are not modified like database_block_siz, but the location and
the patches can be changed
-Oracle database can be started as startup service or use command strdbs.cmd or startdb
(UNIX)
-Oracle startup can be performed using SAPDBA or BR tools
Oracle is started in 3 cases:
*Database is operational when it is started in the open mode
Startups mount:
-It is used to build control files, the system is not usable, and it is started only for
administrative purpose
Startup nomount:
-SGA is built, db views are allowed to access and it is used for recovery purpose
-data files are not opened, db is not usable
Startup Open:
-Database is opened in the open mode and it is accessibility to all users
Oracle is shutdown in 4 cases
Shutdown abort:
-It is used to shutdown the database abruptly, but it is suicidal.
-It should be only used under critical situations
-System is not consistent, so SMON performs redo and undo when the system is restarted
and tries to bring the system to consistent state
-Sometime we may need to restore the system from tapes and recover from logs
-All the opened transactions are abrupted and there is no roll forward or rolled back takes
before shutdown
Shutdown Immediate:
-All opened transactions are rolled back and the system shut downs gracefully, It is
consistent and no restore/ recovery is required
Shutdown transactional:
-All the opened transaction are allowed to complete and the system gracefully, it is
consistent and not restore /recovery is required
Shutdown Normal:
-All the logged users are allowed to work, no new connections are allowed in all the
above cases and the system shutdown gracefully, it is consistent and no restore/recovery
is required
R/3 connection mechanism
Each R/3 work process while starting tries to establish connection to the db
The connectivity as follows
-It loads the dboraslib.dll from run directory and assigns shared memory and get the path
of the oracle home and tries to login to the database as OPS$ users
-In order to connect to database using OPS$ user the 2 parameters needs to be set up
i) remote_os_authnet=true
ii) Os_authnet_prefix=ops$
First, parameter allows, it is used to authenticate os users remotely
Second, when the user is authenticated as admin at os, corresponding user can login to
the database w/o any password
-R/3 workprocess login to the database as OPS$ user, OPS$ user owns only one table i.e.,
sapuers – table consist of SAPSID and its password
-R/3 workprocess get the password of SAPSID user and disconnects from the oracle &
reconnects with user SAP <SID> and known password.
-It is documented in dev_wn log files
* SAP<SID> password resides in 2 tables
SAPUSERS
DBA_USER
-When the password is changes for SAPSID, needs to be checked in 2 locations i.e., the
reason it should be changes in 2 locations and it should be changes using BRtools
>brtools
-4
Lsnrctl
Database is started w/o any error, but sap system could not connect to the database and
results and error ora-1250 (listener is also started)
Each user request is listened on the database on the ports 1521 and 1529
The listener service is running on different ports, other than the ports accessed by r/3 WP
Go to listener.ora and check the listener service is configured for the right port, host
name, sid.
If the port is different, reconfigure the listener service and restart the service
Storage Management
Oracle data is stored interms of data blocks
Data blocks are defined by the parameterd b _ b l o c k _ s i z e
By default it is 8 kb
Oracle data is stored physically on the file structure using data file with extension. d b f,
but they will address logically for easy maintenance and administration, but the data files
with extension. d a t a
SQL.mdf (main data file)
SQL.ldf (log data file)
SQL.ndf (next data file)
Extents
- they are nothing but group of blocks
- they are assigned to a table or index
Extent is a logical container which is assigned to segments (table or index)
While allocating extent to a table the following parameters are used
Initial extent
Next extent
Mini extent
Max extent
Tables
TGORA, IGORA documents the above extents
If a block is opened, it should be written otherwise it is closed and doesn’t open again
Data should not be small
If a block is empty, there is a problem
If data doesn’t fit in 8kb, increase data block to 16kb, otherwise database increases
SAP has 14,000 tables but oracle can hold9 9 9 9 tables , small tables combined to a form
a big tables and the pooled tables and cluster comes into picture
8k is enough for most of the companies
Extent Filling
The tables are populated with the initial extents if the size is known, if not mini extents
are allocated to a table
If mini extent is completed then it will try to grown at the rate of next extent
If the extents are used completely ora 16131 and ora 1632 occurs with a message max
extent reached
If order to adopt next extent schedule SAPDBA –next, BRCONNECT –Fnext
Tables /Indices /Segments
Extents are assigned to segments
A segments is a table or index or a group of a tables
Table space
It is a logical container which consists of tables and index
Each tablespace is referred with a data files on the disk
There are various tablespaces which has their own functionality
There were 27 table spaces up to 4.6c, from 4.7EE there are 4
Prefix
Abbreviation
Ext
PSAP
<TS_name >
D (
R/3 Naming Conventi
1) System table space –It contains the information about the database
2) PSAPTEMP – It is a temporary table space which is used to sort the tables for selected
operations
3) Dictionary Managed Tablespace –this table space managed by the dictionary and it has
its own disadvantage
While allocating, extents contiguous free space required
It cannot coalesce the free space
More disk fragmentation occurs there by increasing the search criteria
Frequently re-organization is required
SAPDBA –next or BRCONNECT –next to be adapted then next extent
4) Locally Managed Table Space –This feature is available since 9.20
This table space has the following features
While allocating next extent, it looks for the free space, and the extents are assigned less
fragmentation occur and frequently and reorganization not required
System defines the next extents SAPDBA –next or BRCONNECT –f next are not
required
BRSPACE provides an option to convert dmts to lmts
5) PSAPROLE-It is a table space which is used to store the data before an update
Datafiles
-These are the physical data files which resides at os level
-Each data file is associated with a tablespace
-A tablespace can have more that 1 datafile, but each datafile can be associated with 1
table space
-Depending on the company category each data file may vary in between 1GB to 8GB
Datafile filling:
Logically user fills the data i.e update the data or inserts the data into tables, which
automatically file the datafile
If the datafile is filled and does not have enough space to insert the data ora 1653, ora
1654 occurs, can be occurred in 2 situations
-When the Datafiles are completely filled up
-unable to extent because there is no contagious space available in the datafile
-The datafile can be resized, > BRCONNECT, space management
Orarchive directory
When ORAARCH directory is filled ORA255 or ORA272 occurs and the systems strucks
and no updates be made
Update gets deactivated, and it should be activated after resolving.
*Resolution: Schedule Archive backup to tape on disk immediately.
Reorganization
Re-organization is required when the db response time are increasing and the database is
growing out of space
The fragmented space is displayed in db02
Process
-Schedule the downtime
-Ensure the SAP reorg die rectory has enough space to hold the size of the db
-Go to SAPDBA or BRTOOLS use the option export
* Ensure that valid backup is available
-Export the database
-Delete the database
-Import the database
BRSPACE provides an option to perform the reorganization online
House-keeping Jobs
DB01 –It is used to monitor the lock waits on the system similar to sm12 on R/3
It helps the lock before a record is updated on the database
DB02 – It is used to identify the space utilization of the tablespace
It is also used to identify the critical objects which are going out of space
It is also used to identify the missing index
Conclusion:
Capacity planning of the db on daily growth
Identify the space critical objects, resize the datafile.
It the tablespace exceeds 90% of the utilization it is consider as critical
Missing indexes:
Missing are displayed using the option missing indexes
Missing indices are displayed in yellow colour select the index and choose the option
create in db or SE14 to create index
BRCONNECT
BRCONNECT –Fcheck: it is used to check the database check condition
BRCONNECT –Fnext: it is used to adapt next extent
BRCONNECT –Fcleanup: it is used the clean the log files
BRCONNECT –Fstats: it is used to update the statistics of all the transparent tables
whose statistics are outdated; it should run weekly atleast once & it is one of the most
important step in installation ,system copy …etc
-identifies the tables whose statistics is outdated and write into the tabled b s t a t c
Update the tables which are in dbstatc
All the above jobs can be performed by db13
DB13
-Double click on date, specify the start time
-Schedule the job
-Execute check and update optimizer statistics
-Adapt next extents
-Check the database
-Verify the database
-Cleanup the logs
ORACLE FOR SAP
BR
BR
Scheduling Databas
Offline backup
Which is used for the environment of 16/5, during the offline backup the database will be
shutdown and table spaces are in the backupmode, it is referred as cold backup
It is consistent to restore
It is scheduled using DB13, BRTOOLs or any third party tools
Offline backup can also include logfiles backups, also referred as complete backup
Online backup
It is used for high availability systems which supports 24x7x365 day-environment
The database should be in archive mode
The users continue to work during the backup operations
It is also referred as hot backup
It is scheduled using db13, sapdba, Brtools
It will be consistent only with the logfiles
Process-
During the backup the tablespace goes into begin backup mode, after the backup the
tablespace goes to end backup mode, during this period the data is written into the log file
Partial Database backup
It is used to backup the database at table space level
Incremental Backup
This is used for the companies where high availability is required and cannot spare much
time for database backup
So the whole database backup is taken wither on Saturday or Sunday, when the whole
backup is taken catalogue is written to the database
From the modified data can be backed up using incremental backup
*Online redolog files:
Orilog A, Mirrorlog A, MirrorLogB B, Orilog B
*Offline redolog files:
The files which are copied from ORAARCH directory to the other files
Redolog Backup
The log files which resides in ORAARCH directory needs to be moved from TAPE
periodically, if they are not moved Orarchive directory fills and archive struck occurs
As the redolog files are critical to restore the database to consistent state, the needs a
backup periodically
If the Redologfiles are lost, the database cannot be stored
Init<sid>.sap:
It is the initialization for Brtools and SAPDBA; it consists of the following parameters
and resides in the database directory oracle\oraa\102\database
BACKUP MODE=all
RESTORE MODE=all
BACKUP TYPE=offline, online
BACKUP_DEV_TYPE=tape or disk
BACKUP_ROOT_DIR=sap directory (backup)
BRARCHIVE_FUNCTION=save
ARCHIVE_FUNCTION=save
ARCHIVE_COPY_DIR=sapbackup\<SID>\Backup
TAPE_COPY_CMD=cpio
TAPE_SIZE
TAPE_ADDRESS=/dev/mnt
VOLUME_ARCHIVE=<specify the name of the tape>
VOLUME_BACKUP=<SID>01
EXPIRE_PERIOD=30 days
TAPE_USE_OUT=<time>
-BRTOOLS takes the parameter from this file by default; however this parameter can be
overwritten by using at run time
BRBACKUP
It is used to backup the database
It can be initialized through SM49
>BRTOOLS – 4
BRARCHIVE
It is used to backup the log files
It can be initialized using SM49
>BRTOOLs -4
BRRESTORE
It is used to restore the system from a database backup
>BRTOOLS -5
BRRECOVER
It is used to recover the database from the log files
BRCONNECT
It is used to establish connection to the database
-It is also used to change the password of SAPSID
Backup database
When a db is backedup its backup all the datafile, control files and log files (tablespace
logically)
BRSPACE
It is the new feature provides by BR tools to perform:
-create, drop, extent, alter a tablespace
-it is also used to alter the database (move)
-BRSPACE can be used interactively to Dictionary managed table space to locally
managed table space
No comments:
Post a Comment