Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
87 Cards in this Set
- Front
- Back
- 3rd side (hint)
The estimation of the hardware requirements necessary to process a workload within specified parameters is referred to as?
|
computer sizing
|
|
|
What are the Parameters necessary to process a workload referred to?
|
Service Level Agreements
|
usually deal with the maximum amount of time allocated to process each workload type.
|
|
What type of requirements does computer sizing estimate?
|
Hardware
|
|
|
What must be defined before you can make computer sizing decisions?
|
Parameters
|
|
|
What Server hardware addition would be needed to decrease Oracle Transaction time?
|
Buying more Disks
|
|
|
What's most frequently needed to reduce Oracle 8 processor utilization?
|
Memory
|
|
|
The performance study of hardware and software resource consumption on an existing system to prepare for the growth of that system's workload is referred to as?
|
Capacity Planning
|
|
|
What is ther term used for projecting Normal growth of CPU utilization; Disk Usage; Memory usage?
|
Capacity planning studies
|
3 words. Ends with "studies"
|
|
What are capacity-planning studies established to maintain?
|
Service level agreements
|
|
|
Why would the activities of specific users be profiled during a capacity-planning study?
|
To predict resource consumption
|
|
|
A database server performs only database functions. In terms of workload, it performs only?
|
Transactions
|
|
|
What are the 2 main types of Transactions?
|
online transaction processing (OLTP) and decision support.
|
|
|
This transaction type is constantly updating the database based on the most current information available so the next user can rely on that information being the most current.
|
OLTP
|
|
|
What's the first category of info gathered in the interview process?
|
concurrent users
|
|
|
What's the second category of info gathered in the interview process?
|
The number of transactions expected to be processed
|
|
|
what's the 3rd type of info gathered in the interview process?
|
what time range these transactions are expected to be processed
|
|
|
what's the 4th type of info gathered in the interview process?
|
Whether this time range is your peak-utilization period
|
|
|
generally used when you cannot define the peak-utilization period.
|
steady-state processing
|
The other philosophy of sizing an OLTP or any other system
|
|
refers to a portion of system processing power left in reserve to accommodate heavier workload periods
|
reserve capacity
|
|
|
This theory states that utilization directly affects the queues and these queues are directly related to response time .
|
knee of the curve
|
the point at which you go from linear growth to exponential growth
|
|
At what point of utilization do queues begin to grow exponentially in length?
|
75%
|
|
|
What principle should you consider when selecting the number of CPU's your system will require?
|
Knee of the Curve
|
|
|
What is the knee of the curve % for disks?
|
85%
|
This percentage applies to the size and I/O capability of the disk drive in question.
|
|
What can be minimized by adhering to the knee of the curve principle as applied to disks?
|
Response times
|
|
|
If a user process needs a data record or a system needs a code page and it is found in memory, this is called a
|
Logical I/O
|
Also referred to as a memory read
|
|
This occurs when the system or a user function determines that certain information that it must process is not in memory?
|
page fault
|
|
|
When the system or user function retrieves the data from physical disk, replacing data currently in memory it is referred to as?
|
physical I/O
|
|
|
Retrieval time for the data found in memory is rated in?
|
microseconds
|
millionths of seconds
|
|
retrieval time for the data found on disk is rated in?
|
milliseconds
|
thousandths of seconds (much longer than memory).
|
|
If a user process needs a data record or a system needs a code page and it is NOT found in memory, what normal system function occurs?
|
page fault
|
|
|
Where does the system retrieve information from during a logical I/O?
|
memory
|
|
|
When the next program code address is not in memory. The system must go to the disk drive to retrieve it. This is called?
|
code page fault
|
This generates a one-way transfer from disk to memory.
|
|
This two-step page fault occurs when a user process requests a data record that is not in memory, and the data page currently in memory has been modified
|
page fault swap
|
|
|
When the data page current in memory has been modified, it's called?
|
dirty data page
|
|
|
Which type of page fault causes the most damage to response time?
|
Page fault swaps
|
|
|
What principle is used in the calculation of minimum memory?
|
page fault principle
|
|
|
This type of system is generally used to provide information to management so decisions can be made about issues such as business growth, levels of stock on hand
|
Decision Support System
|
also referred to as DSS
|
|
What is most important in DSS Systems
|
minimize process request time
|
With a DSS system, keeping processor utilization under 75% is less important
|
|
In the OLTP system, the rate of the throughput of transactions is commonly measured in
|
tps or tpm
|
|
|
With DSS, throughput is usually measured in?
|
qph
|
|
|
This system can be 1 TB in size and Measures throughput in QPH
|
DSS
|
|
|
This system measures throughput in TPM and must have 75% processor utilization?
|
oltp
|
|
|
3 factors for determining cpu and memory requirements are?
|
number of users, transaction type and OS
|
|
|
To size a database server you need to know which 2 things?
|
number of concurrent users and number of transaction i/o user requests will generate
|
|
|
When calculating the proper amount of memory for your system, the person performing the sizing must take into account?
|
desirable cache-hit rate and page faulting
|
|
|
What information do you need to size a database server?
|
Number of concurrent user connections and number of transaction I/Os
|
|
|
Oracle 8 uses how much memory for each user connected?
|
500kb
|
|
|
What is the memory usage for Oracle on Windows NT?
|
18mb
|
|
|
Oracle 8 database executable used how much memory?
|
5.5mb
|
|
|
What information needed to size a database server is indicative only to Oracle databases?
|
size of SGA
|
It provides for variable cache sizes, allowing for high cache-hit rates and high-speed processing.
|
|
What are SGA's 3 main facets that should be considered for determining Oracle 8 memory needs?
|
shared pool size, log buffer size and database buffer size
|
|
|
SGA shared pool area consists of these 2 sub areas?
|
library cache and dictionary cache
|
The shared pool contains such things as stored procedures, stored functions, and SQL statements.
|
|
contains procedures that are internally formatted (parsed) and are commonly used by the system
|
library cache
|
The effect that you, the system administrator or performance technician, want is for a procedure to be parsed once, but used multiple times.
|
|
This cache contains the most frequent requests pertaining to the database.
|
dictionary cache
|
A validation of access privileges is an example of this type of request.
|
|
Shared pool memory size for a)<44 users, b)44-95 users, c)95+ users
|
a)3.5mb, b)6mb, c)9mb
|
|
|
What does the dictionary cache contain?
|
database requests
|
|
|
the second part of the SGA that determines the size of memory and holds information about the write activity taking place.
|
log buffer area
|
contains information showing the changes that have been made to the database
|
|
recommended log buffer size for a)<44, b)44-95, c)95+
|
a)8,192bytes, b)32,768bytes, c)163,840bytes
|
|
|
When the log buffer fills up, this process writes the contents of the buffer to the redo log files associated with the database.
|
lgwr
|
If this buffer is too small, this process activates more often, thus slowing the performance of your system because of the write activity to the files
|
|
This SGA area is the cache area of your system. All the data that is processed by your system will pass through this area.
|
database buffer
|
|
|
What size does the calculation for log buffer size relate to?
|
Concurrent user community
|
|
|
Within what file does the parameter DB_BLOCK_BUFFERS exist?
|
initorcl.ora
|
suggests the values that are a recommended starting point for the DB_BLOCK_BUFFERS parameters.
|
|
refers to the quantity of database buffers that this system will have?
|
DB_BLOCK_BUFFERS
|
specified in the initorcl.ora file
|
|
refers to the size that each block will be
|
DB_BLOCK_SIZE
|
specified in the initorcl.ora file
|
|
Which two things increase when a cache area is too small?
|
occurrence of physical i/o's and response time.
|
|
|
What increases relative to how high the DB_BLOCK_BUFFER parameter is set?
|
transaction completions
|
|
|
The target cache-hit rate for your system should be as close to what percent?
|
90%
|
|
|
What do you need to calculate once you know how to set DB_BLOCK_BUFFERS properly?
|
DB_BUFFER_SIZE
|
|
|
These two scripts may use to retrieve additional information needed to calculate the target cache-hit rate. script a) before performing test run and b) after test run.
|
a)utlbstat.sql and b)utlestat.sql
|
|
|
What script is used to calculate the target cache-hit rate?
|
ulestat
|
|
|
What does understanding the transaction anatomy help calculate?
|
CPU utilization
|
|
|
allow the system to split the access of the mirrored drives in two.
|
split seaks
|
This greatly accelerates the time it takes to locate the desired data and thereby reduces transaction response time.
|
|
how to determine i/o for RAID 1?
|
I/O=READS_PER_TRANS+(WRITES_PER_TRANS*2)
|
|
|
How to determine RAID 5 I/O total?
|
RAID 5 Trans I/O=(READS_PER_TRANS+(WRITES_PER_TRANS*2 for reads))+(WRITES_PER_TRANS*2 for writes)
|
|
|
What must be calculated before figuring anticipated CPU utilization?
|
total reads
|
|
|
What type of reads involve more intervention by the CPU?
|
physical
|
|
|
allows a large query to be broken into segments based and to hit the database simultaneously, thus expediting processing.
|
Parallel Query
|
|
|
Approximately how many seconds does it take to write a block from memory to disk?
|
.001
|
|
|
What percentage rate should the number of proposed CPUs keep TOTAL_CPU_UTILIZATION below?
|
75%
|
|
|
What order should these be in to keep cpu utilization under 75% 1)Calculate Total Writes 2)Calculate Transaction Utilization 3)Calculate CPU Write Time
|
132
|
|
|
What sizing option is available on a DSS system utilizing multiple CPU's?
|
parallel query
|
|
|
what is the disk usage percentage when sizing the disk subsystem?
|
85%
|
|
|
What does OS_ORACLE_DISK*RAID_FACTOR determine
|
Number of disk drives
|
|
|
what raid level is recommended for OS and Oracle Database?
|
Raid 1
|
you will want the fastest possible recovery for the operating system and the Oracle executable.
|
|
what RAID level is recommended for the database drives?
|
RAID 5
|
|
|
What kind of information is kept on log disk drives?
|
Audit Trails
|
|
|
What variable is extremely important when calculating the number of log disk drives?
|
Number of writes-per-second
|
|