Archive

Archive for the ‘Servers’ Category

Windows 2008 Power Schemes and Performance Impact

July 11, 2011 1 comment

What is a Power Plan ?

A power plan is a collection of hardware and system settings that manage how your computer uses power. You can use power plans to reduce the amount of power your computer uses, maximize performance, or balance e two.

Power plans available with Windows Server 2008 are Balanced , Power Saver & High performance.

In Microsoft Windows 2008  &  Windows 2008 R2 Operating Systems , the default Power Scheme is set to be as Balanced which gives a balance between Performance and Energy Savings.

Doing some research on the impact of Power Scheme on Performance of Windows 2008 Hosts and specifically Hyper-V Hosts led me to multiple articles that has been referred below. Interestingly there is a well documented Microsoft Knowledge base article as well that you can find here. Although the KB does not say how much of an impact can the change have , there are some bloggers who have shared their testing result on the same.

Changing default power scheme from “Balanced” to “High Performance” has yielded 22% improvements in SQL performance that has documented by Glen out  here and Improvements in Hyper-V Performance can be found here. In our preliminary testing , we are seeing better results for Performance of  Hyper-V hosts that has been changed from default “Balanced” Power Scheme to “High Performance” Power scheme.

For people who are interested in Changing the Power scheme –

In a Full Installation of Windows Server 2008 , Open Control Panel –> Power options and you will be bale to select the Power Scheme of your choice

In a Core install of Windows Server 2008 , One can make use of powercfg.exe command to change the power scheme.

” powercfg -getactivescheme ” would dispplay the active Power Scheme in the Host

” powercfg -setactive 8c5e7fda-e8bf-4a96-9a85-a6e23a8c635c ” use this command to set the Power Scheme to High Performance

We are still doing our round of  testing to validate the improvements in Performance of Hyper-V hosts and Windows 2008 hosts due to Power Scheme. Would love to your experience and feedback as well.

Advertisements

Verifying File Integrity using MD5 Checksum

May 22, 2011 1 comment

What is MD5 Checksum ?

According to Wikipedia ,  MD5 Message-Digest Algorithm is a widely used cryptographic hash function with a 128-bit (16-byte) hash value. Specified in RFC 1321, MD5 has been employed in a wide variety of security applications, and is also commonly used to check data integrity.An MD5 hash is typically expressed as a 32-digithexadecimal number. MD5 is more commonly used these days and is used to ensure the file integrity of files that we download.

How to verify MD5 for files that we download ?

Even though there are many tools to verify MD5 , I prefer the Microsoft FCIV (File Checksum Integrity Verifier) Utility which can be download from here.Fciv is a command line utility that computes and verifies hashes of files.It computes a MD5 or SHA1 cryptographic hash of the content of the file.

The FCIV utility has the following features:

  • Supports MD5 or SHA1 hash algorithms (The default is MD5.)
  • Can output hash values to the console or store the hash value and file name in an XML file
  • Can recursively generate hash values for all files in a directory and in all subdirectories (for example, fciv.exe c:\ -r)
  • Supplies an exception list to specify files or directories to hash
  • Can store hash values for a file with or without the full path of the file
Let’s try FCIV Utility to verify MD5 & SHA1 SUM for a vSphere Hypervisor Download.
Setting up FCIV Utility :
  1. Download FCIV Utility “Windows-KB841290-x86-ENU.exe
  2. Double click Windows-KB841290-x86-ENU.exe and extract it to a folder
  3. Add the FCIV folder to the system path or else every time you execute FCIV, one has to browse to the folder where  FCIV Utility has been  extracted .
  4. Go to command prompt and type fciv.exe /?, and then press ENTER to see the various options.
Ensuring File Integrity would mean that MD5 or SHA1 SUM value displayed on the website should match with that of the downloaded file.
In our case , VMware ESXi 4.1  download page tells me the MD5 and SHA1 SUM value for the ESXi 4.1 Update1 ISO image.
Once the download is complete , Open a command prompt and calculate the MD5 or SHA1 SUM as shown below using the FCIV utility .
fciv -add  -md5
fciv -add  -sha1
Categories: Free Tools, Servers Tags:

Post P2V Tasks

January 9, 2011 1 comment

P2V does not end with migrating a Physical Server to Virtual Server  but there are certain post P2V tasks that could help in improving the performance of the converted server. 

  • Ensure you install VMware Tools on the migrated VM without fail.
  • Remove all unwanted hardware components from a migrated virtual server like serial ports and floppy drives.
  • Change the CPU and memory of the VM as per requirement and do not go with default values as they reflect the resources as in the Physical Server.
  • Physical Servers are installed with OEM drivers which are resource consuming and these drivers can be removed using Add/Remove Programs once they have been converted to VMs.

I came across a HP utility that can be download from here  which can be used to remove HP PSP with a single click.

One of the issue that is reported is the error when they get when try to assign IP to the migrated server –  ” The IP address XXX.XXX.XXX.XXX…” is already assigned to another adapter”.

This error message is caused by Ghosted NIC Cards that are present. These are the NIC cards that are converted as part of the P2V process since they were part of the physical Server but does not show up on the device manager. Follow the below mentioned steps to uninstall the ghost NIC cards.

  • Click Start –>Run–>Cmd [ Open a Command prompt]
  • Type in “set devmgr_show_nonpresent_devices=1” [ This change would display all non-present devices in device manager]
  • From the command prompt, type in “devmgmt.msc” [Do not go to Run prompt to open device manager or open from a GUI since above change would be effective from cmd prompt only]

            

  • Click View –>Show Hidden Devices
  • Non-Present devices are grayed out and can be uninstalled from the server. In this case  out NIC cards has to be uninstalled.

             

  • Now try assigning the IP back to the NIC card in the Virtual Server and should work fine.

IP Address issue has been addressed by Microsot here and VMware here and lot of other bloggers too.

Understanding RAID Controller and RAID Performance

July 28, 2010 Leave a comment

Server can be provisioned with disk space from either Storage ( SAN , NAS , NFS etc .. ) or DAS ( Direct Attached Storage ) . There are lot of posts from the Storage Vendors on the Performance considerations that needs to be accounted in a Storage environment to achieve desired level of performance but i have not come across many articles on Performance considerations for DAS Environments . This blog post will cover some basic factors that has to be considered in a DAS environment .

RAID – Redundant Array of Disks are used in Servers and Storage to server three main purpose

  1. Increasing Reliability and providing Fault Tolerance
  2. Creation of large storage Volumes
  3. Increasing I/O capabilities and throughput

In Storage Systems , RAID is taken care by the Storage System whereas Servers come with RAID Controllers which help in providing RAID capabilities to local disks .

Overall Performance of RAID arrays in Servers are dependent of the below factors

  • RAID Controller Performance
  • RAID Level
  • Number of Disks and Drive Performance
  • Storage Interface Performance ( SAS vs SATA )

I will stick to explaining briefly understanding RAID Controller performance in this post as the other three factors are same as in Storage environments .

RAID Controller Performance :

Two important factors in a RAID controller that contributes to the performance of the RAID controller  are Processor and the associated Read / Write cache . Processor is responsible for processing all RAID operations such as write operations for redundant RAID modes  and calculating Parity bit that is written onto disk based on RAID level which in future would be used in rebuilding of data  based on parity in case of drive failures .

Cache :

Amount of Cache available on RAID controller varies with vendors and we will have to optimally split the cache into Read Cache and Write Cache  based on our requirements. Read Cache are helpful in scenarios where the data is read sequentially . RAID controllers will pre-fetch data from the disk the moment it detects sequential data and store it in the cache which will in turn mean faster response time .

Read Cache might not be effective in scenarios where the Reads are random . Also , Please note that the read performance of the disk arrays would be far better than writes ( due to associated disk penalties )  and hence most RAID controllers are configured for 25%:75% ( Read Cache:Write Cache).

As in Storage environments , Write cache are used as buffers wherein the host will write data onto the cache and acknowledgement will be sent from the cache . RAID controller will later write the data from Cache onto Disks .This helps in achieving higher disk throughput . Various techniques write like write-coalescing and command reordering are employed  to write data back to disk effectively .

Note that the acknowledgement is sent from the cache and it is imperative to write back the data to the disk to maintain data integrity . To maintain this integrity in cases of sudden power failure or a server crash , cache is generally backed by battery and it is advisable not to use cache in cases where there is no battery backed cache .

Categories: Servers Tags: , ,

Introduction to UCS – Unified Computing System

July 17, 2010 1 comment

Recently I attended a session on UCS and was completely amazed at the amount of work put in by CISCO folks on the UCS and hence i decided to briefly explain the various components of UCS and why it is such an amazing product .  True to its name , CISCO’s UCS is a complete unified system that addresses some limitations with Servers that are available in the market today . UCS Blade Servers combined with the management software UCS Manager makes a great product. UCS seems to have been inspired from VMware in making the product Scalable , Manageable and Reliable. UCS is made for the Enterprise and purposely built for Virtualization . A strong 10GE and SAN infrastructure would make UCS an invaluable asset to any Organization .

UCS is available in two variants :

  • B Series – Blade Servers
  • C Series – Rack Mount Servers

B Series Blade Servers  :

UCS B Series is made up of Blade Chassis that consists of Blades and Fabric Extenders that connects to a Fabric Switch with Expansion Modules which in turn connects to the Corporate Network. All UCS Blade chassis and its components can be managed from UCS Manager .

UCS Manager :

  • Provides unified, embedded management of all software and hardware components (  Adapters,Blades,chassis , fabric Extenders and Fabric Interconnects ) of the Cisco Unified Computing System – across multiple chassis and thousands of virtual machines.
  • UCS Manager can be used to Discover , Configure , Inventory , Monitor , Diagnose and collect statistics of all the components.
  • CLI can also be used to manage the chassis and its components
  • OS Installations can be carried out from UCS Manager
  • KVM Consoles for the blades are available through UCS Manager

Blade Chassis & Its Components :

UCS consists of a 6U Blade Chassis with a standard front to back cooling that is scalable and flexible . This chassis can accommodate 8 Half Height Blades or 4 Full Height blades or a combination of half and full height blades .

Half Height variants that are available are :

Cisco UCS B200 M1 – Intel Xeon 5500 series powered dual processor with 12 DIMM slots that supports upto 96GB RAM and consists of 2 Internal 2.5″” SAS HDDs with an integrated RAID  support for RAID 0 and RAID 1

Cisco UCS B200 M2 – Intel Xeon 5600 series powered dual processor with 12 DIMM slots that supports upto 96GB RAM and consists of 2 Internal 2.5″” SAS HDDs with an integrated RAID  support for RAID 0 and RAID 1

Full Height Blades variants that are available are

Cisco UCS B250 M1 Extended Memory Blade Server – Intel Xeon 5500 series powered dual processor with 48 DIMM slots that supports upto 384GB RAM and consists of 2 Internal 2.5″” SAS HDDs with an integrated RAID  support for RAID 0 and RAID 1

Cisco UCS B250 M2 Extended Memory Blade Server – Intel Xeon 5600 series powered dual processor with 48 DIMM slots that supports upto 96GB RAM and consists of 2 Internal 2.5″” SAS HDDs with an integrated RAID  support for RAID 0 and RAID 1

Cisco UCS B440 M1  High Performance Blade Server – Intel Xeon 7500 series powered Quad processor with 32 DIMM slots that supports upto 256GB RAM and consists of 4 Internal 2.5″” SAS HDDs with an integrated RAID  support for RAID 0 , 1 , 5 and RAID 6 .

Fabric Switch is available in 20 Port and 40 Port configuration .

Expansion Modules is available in FC only ( 8 SFP Ports that run 1,2 & 4 Gbps FC ) , Combo FC + Ethernet  ( 4 SFP+ Ports that run 10GE &  4 SFP Ports that run 1,2 & 4 Gbps FC ) or Ethernet only ( 6 SFP+ ports that run 10GE )

Service Profiles

  • Service Profiles stores a server personality which includes UUID of the Server , identity of NIC and HBA , Boot order and various policies and what Service Profiles brings to UCS is Server Mobility .
  • Service Profiles can be disassociated and associated with any blade across any UCS Chassis. This helps administrators move Applications between blades provided OS and Data Resided on the SAN
  • Service Profiles can be cloned or deployed from Templates . Service profiles are associated with blades and can be moved between various blades .
  • Auto populating Server Pool is another cool feature wherein administrators can create an empty server pool and map it to a Server pool qualification specifying various qualification that could qualify a server for the requirement and the server would automatically be placed in the Pool

Monitoring and Reporting

  • We can configure call home to automatically notify CISCO or Administrators in case of any faults with the UCS Blades
  • Upto 3 Syslog Server can be configured to log events associated with the servers
  • SNMP is supported
  • Threshold policies can be set to trigger alarms based on conditions
  • inventory reports can be exported
  • Statistics is used to display statistics of almost all components that are part of UCS

Some advantages of UCS are

  • Centralised monitoring and reporting of all components of UCS
  • Move Applications across various blades using Server Profiles
  • Virtualization ready
  • Server Mobility
  • Diskless Configuration i.e SAN Boot is completely supported
  • Scale upto 40 Chassis and 320 Computer nodes
  • Granular Delegated Management is possible

To Conclude I believe UCS would be a great asset in environments with a Strong 1oGE and SAN infrastructure . UCS Manager is a great piece of management software that could help in managing and reporting centrally but will need some time for the administrators to get used to .

For more on UCS ,

Comparison between UCS and HP Blade : http://www.bythebell.com/2010/07/cisco-ucs-vs-hp-bladesystem-matrix-an-update.html

All Things UCS : http://viewyonder.com/cisco-ucs/