How PERFMON (Windows tool) helps DBA to identify the Performance Issues and take necessary solutions

In SQL Server DBA Environment, there are several ways to improve the performance of SQL Server Database Applications such as Query Execution Plans, Sql Profiler, DTA (Database Tuning Advisor), Server Level properties, Object level improvements (Indexes, statistics, other maintenance stuff) and Windows Level Applications. I mean the database performance can be identified in 3 levels i.e., Server Level (Operating Systems, Networking Protocols); Database Level (Sql Server Databsae Engine, SQL Server) ; Object Level (Objects within the database).


So if we are facing the performance degrade then we need to check up at all levels for taking necessary actions. Currently I will discuss about the Server Level Tool PERFMON which is very useful to identify the sever levels activities, based on the identified results we can take appropriate action to improve the performance. It can be Hardware, Memory (or) System changes.


PERFMON : PERFMON is a windows inbuilt tool which can provide the workload of the resources running in the system. It can be used to find out Windows resources data as well as SQL Server resources.


Main Benefits Of The Tool :
• Understand your workload and its effect on your system's resources.
• Observe changes and trends in workloads and resource usage so you can plan for future upgrades.
• Test configuration changes or other tuning efforts by monitoring the results.


System Monitor and Performance Logs and Alerts provide detailed data about the resources used by specific components of the operating system and by programs that have been designed to collect performance data.

Choosing the data to monitor :
Start by monitoring the activity of the following components in order:
• Memory
• Processors
• Disks
• Network


Following counters can be helpful to trace the data

1:
Component : Disk
Performance aspect being monitored : Usage
Counters to monitor :
Physical Disk\Disk Reads/sec, Physical Disk\Disk Writes/sec, LogicalDisk\% Free Space, Interpret the % Disk Time counter carefully. Because the _Total instance of this counter may not accurately reflect utilization on multiple-disk systems, it is important to use the % Idle Time counter as well. Note that these counters cannot display a value exceeding 100%.


2 :
Component : Disk
Performance aspect being monitored : Hindrances
Counters to Monitor : Physical Disk\Avg. Disk Queue Length (all instances)


3:
Component : Memory
Performance aspect being monitored : Usage
Counters to Monitor : Memory\Available Bytes, Memory\Cache Bytes


4:
Component : Memory
Performance aspect being monitored : Hindrances
Counters to Monitor : Memory\Pages/sec, Memory\Page Reads/sec, Memory\Transition Faults/sec, Memory\Pool Paged Bytes, Memory\Pool Nonpaged Bytes.
Although not specifically Memory object counters, the following are also useful for memory analysis: Paging File\% Usage object (all instances), Cache\Data Map Hits %, Server\Pool Paged Bytes and Server\Pool Nonpaged Bytes


5:
Component : Network
Performance aspect being monitored : Throughput
Counters to Monitor : Protocol transmission counters (varies with networking protocol); for TCP/IP: Network Interface\Bytes total/sec, Network Interface\ Packets/sec, Server\Bytes Total/sec, or Server\Bytes Transmitted/sec and Server\Bytes Received/sec


6:
Component : Processor
Performance aspect being monitored : Usage
Counters to Monitor : Processor\% Processor Time (all instances)


7:
Component : Processor
Performance aspect being monitored : Hindrances
Counters to Monitor : System\Processor Queue Length (all instances),
Processor\ Interrupts/sec, System\Context switches/sec



How to Create and perform :
1. Go to RUN and type PERFMON then Enter
2. Double-click Performance Logs and Alerts, and then double-click Counter Logs. Any existing logs will be listed in the details pane. A green icon indicates that a log is running; a red icon indicates that a log has been stopped.
3. Right-click a blank area of the details pane, and click New Log Settings.
4. In Name, type the name of the log, and then click OK.
5. On the General tab, click Add Objects and select the performance objects you want to add, or click Add Counters to select the individual counters you want to log.


Configure the other setings as you required. It can be run for certain time period i.e, 10 hours, 1 day, 1 week to identify how the processes are running in the system. The data can be saved as .CSV or text files. When you get the data you can make a charts using Excel and it can be understandable what necessary action to be taken for improving performance.

Partition a Table for better performance in Sql Server 2005 / 2008

In earlier versions we didn't have partitioning a table option to distribute the data across multiple file groups. This was introduced with Sql server 2005 including other valuable features. Partitioning can be performed with help of Partition Function, Partition Scheme.

Partition is nothing like distributing the data across multiple file groups, as we all know if we share the data into multiple drives automatically performance of the database will be increased. Let assume we have a big table comprises million of records, when fetching the data it took too long time. So, for increasing the performance of the table we can implement Partitioning concept and achieve the better performance.

HOW TO PARTITION A TABLE :
Requirement : A table is to be defined with multiple columns, out of which we need to fetch the data based on column "Name". If you select data based on Name there would be many records with different names comprising letters from A-Z. So by using following method, the data can be distributed into different file groups.

Step 1 : As a first step you need to create a Partition function :
CREATE PARTITION FUNCTION NameFunc (nvarchar(30))
AS RANGE RIGHT FOR VALUES ('H', 'N')


Step 2 : Need to create partition scheme based on the partition function :

CREATE PARTITION SCHEME NameScheme
AS PARTITION NameFunc TO (fg1, fg2, fg3)


Step 3 : Create a table based on Function and Scheme

CREATE TABLE dbo.Employees
(EmpID int, Name nvarchar(30))
ON NameScheme (Name)



Explanation :
In the first step 1, 'H', 'N' values defines
A-H first group,
H-N second group,
N-Z Third group

the above 3 groups will be stored on fg1, fg2 and fg3 file groups respectively.