SAN-BOOKzv

1

INDEX1. Raid--------------------------------------------------------------------------------------------03 (I)Different types of Raid models: R0, R1, R3, R5, R6, R10, R01----------042. Storage Areas -------------------------------------------------------------------------------08 (i) DAS (Ii NAS (iii) SAN3. Difference between File level I/O & Block Level I/O-------------------------094. Fiber channel Topologies -------------------------------------------------------------10 (I)Point-to-Point Topology (ii) Arbitrated loop Topology (iii) Switched Fabric Topology5. Protocols: -----------------------------------------------------------------------------------12 (i) SCSI (ii) iSCSI (iii) FCP (iv) iFCP (V) FCIP (vi)FCOE6. FCLayers: ------------------------------------------------------------------------------------15 FC0, FC1.FC2, FC3, FC47.Disks (i)ATA (ii)SATA (iii)SCSI (iv)iSCSI (v)FCD (vi)SSD-----------------------17

EMC1. Array models ---------------------------------------------------------------------------- --- 21 2. Architectures of CX-300, CX-500 & CX-700----------------------------------- 223. Difference between CX-300, CX-500 & CX-700------------------------------- --24 4.Difference between CX3-10, CX3-20, CX3-40, & CX3-80 --------------------25 5.Difference between CX4-120, CX4-240, CX4-480 & CX-960----------------256. FLARE---------------------------------------------------------------------------------------- --26 7.How to install Flare-------------------------------------------------------------------------27

1

8. NDU (NON-DISRUPTIVE-UPGRADE) ------------------------------------------------289. Command Line Interface (a) Navi CLI (b) Navi Secure CLI-------------------2910. Layered Applications--------------------------------------------------------------------30 (I)Snap view----- (a) Snap Shot (b) Snap clone (ii)Mirror view--(a) Mirror view synchronous (b) Mirror view Asynchronous (iii) San copy--- (a) Incremental copy (b) Full copy11. HBA ------------------------------------------------------------------------------------------ 32 12.Power path ----------------------------------------------------------------------------------33 13.Power path policies-----------------------------------------------------------------------34

HP1.Introduction ------------------------------------------------------------------------------- -36 2.Array models----------------------------------------------------------------------------------36 3.Architectures of EVA 4400, EVA 6400 & EVA 8400----------------------------37 4.HP Enclosures ----------------------------------------------------------------------------- --38 5.Disk group -------------------------------------------------------------------------------- ---39 6.Layered Applications ---------------------------------------------------------------------41 17.RSS------------------------------------------------------------------------------------------- -41 8.Load balancing policies in HP----------------------------------------------------------42

SWITCHS 1.Introduction -------------------------------------------------------------------------- ---- 44 2.Different types of ports -------------------------------------------------------------- ---45 3.F-Logi & P-Logi------------------------------------------------------------------------------46

1

4. Zoning----------------------------------------------------------------------------------------47 (a) Hard zoning (b) Soft zoning

VM-WARE 1.Defination ----------------------------------------------------------------------------------49 2.Virtualization we have two servers ----------------------------------------------- 49 (a) GSX Server (b) ESX Server 3.Two file systems--------------------------------------------------------------------------50 (a) RDM (b) VMFS 4.Two software’s to connect ESX Server’s-----------------------------------------50 (a) VMIC (b) VMIS 5.V-Motion-------------------------------------------------------------------------------------50

1

RAID

RAID (Redundant Array of Independent Disk/Inexpensive Disk):

Raid is an array of multiple independent hard disk drives that provides high performance more power and fault tolerance.

Advantages:

Redundancy of Data (No Data Loss)

If there is any disk failure, the data on that disk is still available with the concept of redundancy, using the mirroring and parity concept the data that is lost is retrieved back again. Thus we maintain the availability of data all the time.

Availability of Data is High & Fault Tolerance

Ex. In spite of any failure of a component (disk) in the process of IOPs, the data must be available for the I/O processing .Using the parity or mirroring and hot spare concepts the IPOs works non disruptive. Thus data availability is high with raid concept.

RAID GROUP: The main objective of raid group is to combine different disks and make it as a single Row entity and improving storage space.

Different types of Raid Models:

R0, R1, R3, R5, R6, R10, R01

R 10 and R 01: Both are called Hybrid Raid Levels

1

RAID 0:

The data will be striped across the disk 100% storage space utilized Performance is High, Write and Read Performance is High

Disadvantages: In this level if one disk is failed, all the data will be loss Min 2 disks & Max 16 disks• Not a "True" RAID because it is NOT fault-tolerant• The failure of just one drive will result in all data in an array

being lost

RAID 1:

Mirroring concept can be used in RAID 1 Level In case ,if any disk is failed, By using Mirror disk we can rebuild

the data Here Write performance is Medium but Read performance is

High Space utilization is 50% only(disk space utilization) Min it will support 2 disks & Max 16 disks 16 disks means 8 Mirror pairs

Advantages:

• Twice the Read transaction rate of single disks, same Write transaction rate as single disks

• Simple to implement• Highest level of Fault-tolerant• Requires a minimum of 2 drives to implement

Disadvantage:

1

Requires twice the disk capacity of required space

RAID 3:

Min it will support 3 disks & Max 16 disks Last disk will go for the parity. It means user data will not be

stored in the last disk Dedicated parity can be achieved by performing EX-OR

operation It means if any one disk is failed in an array, it will rebuild the

data by using parity Striping with byte level In RAID 3 we can use 70% space only(n-1 disks)

Disadvantage:

• This technology is fairly complex. • Performance is slower for random, small I/O operations.• Speed is limited by the slowest disk

RAID 5:

This level is very commonly used and is important Data will striped across the disk It’s having a parity called as Diagonal parity Striping with block level The write performance is bit slow compare to RAID 3 Read performance is High Min it will support 2 disks & Max 16 disks

Disadvantage:

• Disk failure has a medium impact on throughput

1

• Most complex controller and its Firmware design - Difficult to rebuild in the event of a disk failure

• Medium Write data transaction rate

RAID 6:

It will require Min 4 disks ,Max 16 disks It has Row parity(Dedicated parity) and Diagonal parity It supports only Even no. of disks If we loss 2 disks at a time there is no impact on the data I/O operations are High

Disadvantage:

• More complex controller and its Firmware design• Slow write performance because of Controller overhead to

compute parities • Requires N+2 drives to implement because of dual parity scheme

RAID 10(0+1):

It Mirrors the data first and then stripes that Mirrored data into disks

It will require Min 4 disks ,Max 16 disks High availability of data It has High I/O Rates

RAID 01(1+0):

It stripes the data and then Mirrored the data into disks It will require Min 4 disks ,Max 16 disks

1

High availability of data It has High I/O Rates

JBOD (Just Bunch of Disks):

In this we can club 0-14 disks It will support RAID 1,RAID 3,RAID

RAID Raid 0 Raid 1 Raid 3 Raid 5 Raid 6 Raid 10(0+1)

Raid 01 (1+0)

Minimum No. of disks required

2 2 3 3 4 4 4

Maximum No. of disks required

16 16 16 16 16 16 16

Space utilization

100% 50% 70% More than 70%

60% to 70%

Less than 50%

Less than 50%

PerformanceReadWrite

HighHigh

HighMedium

HighHigh

HighMedium

HighHigh

HighHigh

HighHigh

Data Redundancy

No Yes Yes Yes Yes Yes Yes

Striping Block level

Block level

Byte level Block level

Block level

Block level

Block level

Techniques Striping Mirroring

Dedicated parity

Diagonal parity

Dual parity

Mirroring then Striping

Striping then Mirroring

Disks used Any no. of disks

Only even no. of disks

Any no. of disks

Any no. of disks

Any no. of disks

Only even no. of disks

Only even no. of disks

STORAGE ARRAYS:

There are three types of storage arrays:

1

DAS: Direct Attached Storage

NAS: Network Attached Storage

SAN: Storage Area Network

NOTE:

CIFS: Common Internet File System NFS: Network File System

DAS (Direct Attached Storage):

The name itself says that there is a direct connection b/w Host & Array

It is very simple to configure based on Array configuration We can connect Max 16 servers to an Array It supports a Block Level I/O(It access the data in Block Level) In this we use Fiber channel(FC) protocol It has the limited connectivity of servers based on Array model In FC, the data will be transferred in the form of frames The size of each frame is 2 MB I/O transfer rate is Max 8Gbps, Min 1Gbps

NAS (NETWORK ATTACHED STORAGE):

In NAS, it is a File Level I/O Data will transfer in the form of packets from the Host to NAS

box and From NAS box to Array the data will transfer in the form of Frames

Data transfer rate is slow(<8Gbps) when compare to DAS

1

We use CIFS(If it is WINDOWS OS) & NFS(For UNIX OS) protocols NAS box is the Manager b/w the LAN switch & Array

SAN (STORAGE AREA NETWORK):

A Storage Area Network is a dedicated storage network intended to optimize any-to-any communication between network servers and disk arrays. It allows multiple servers of different types and platforms to access the same storage.

It is a Heterogeneous Environment Each and Every thing is a clustered component, clustering means

High availability SAN is a Fabric based FC protocol is used Data is transferred in the form of Frames Max 15 million nodes we can connect It’s support Block Level I/O(It access the data in Block Level I/O)

Advantages:

High Availability of data Performance is High Centralized data storage in the form of Frames

Difference b/w Block Level I/O & File Level I/O:

In SAN , It access the data on Block Level and produces the space to Host in form of disk

In NAS, It access the data on File Level and produces the space to Host in form of N/W folders

DAS NAS SANNo. of devices 16 126 15 MillionData transfer Frames First packet till NAS Frames

1

BOX then FramesSpeed 8 Gbps Less than 8 Gbps 8 GbpsI/O’s Block Level File Level Block LevelProtocols Fiber channel

ProtocolCIFS,NFS Fiber channel

Connectivity Fiber channel Until NAS BOX “Ethernet” cables and from NAS BOX FC to array

Fiber cable

Cost of Maintenance High Low compare to DAS High

FIBER CHANNEL TOPOLOGIES:

There are three major Fiber Channel Topologies:

1) Point-to-Point Topology2) Arbitrated Loop Topology3) Switched Fabric Topology

(1)Point-to-Point Topology:

It is a direct connectivity between Host and Array. Data transfer rate is high. This is a simplest topology with limited connectivity.

Disadvantages:

It has limited connectivity It will support Maximum 16 devices we can connect

(2)Arbitrated Loop Topology:

In all devices are in loop or ring structure. Adding or removing a device from the loop causes all activity on

loop to be interrupted

1

The failure of one device causes a break in the ring. Due to above reason we go for FC HUB concept, in that all the

devices are connected to FC HUB, if suppose any one device fails, simply we can remove that device and then remaining all devices are working fine in network.

Disadvantages:

Performance is slow because Here we are using FC HUB that means main functionality of HUB is, if any Host sends the request then the HUB will broad cast the request to all the Hosts which are connected to the HUB

It will support Maximum 126 devices we can connect

(3)Switched Fabric Topology:

Switched fabric means interconnect of switches together is called switched fabric or [Two or More switches together is called switched fabric]

In a single switched Fabric Maximum we can connect 239 switches

By using this Switched Fabric we can connect Max 15 Million Nodes

Performance is High Most of the people will use the Switched Fabric

PROTOCOLS:

1

PROTOCOL: We follow some set of rules which is used to communicate from one end to another end.

There are Different types of protocols:

1. SCSI: Small Computer System Interface2. iSCSI: Internet Small Computer System Interface3. FCP: Fiber Channel Protocol4. iFCP: Internet Fiber Channel Protocol5. FCIP: Fiber Channel Over Internet Protocol6. FCOE: Fiber Channel Over Ethernet

(1)SCSI:

It is a basic protocol SCSI Protocol which can directly communicate within the devices SCSI protocol works with a concept of Initiator and Target Initiator means always sends the request to Target and Target

will responds to the Initiator’s request Initiator is nothing but Host SCSI has its own commands Data transfer will be in the form of CDB(Command Descriptive

Block) CDB contains the Initiator information, target information along

with the SCSI commands

Disadvantages:

SCSI has distance limitations

(2) iSCSI:

1

It is nothing but Transporter of SCSI protocol through the internet

iSCSI protocol works with a concept of Ethernet Technology Now-a-days it is very useful and importance It has its own processor but LAN doesn’t have its own processor

it depends upon system processor Data transfer will be in the form of PDU’s(Protocol Data Units) PDU internally contains payload and CDB Payload contains the information of data iSCSI is important to SAN technology because it enables a SAN to

be deployed in a LAN,WAN or MAN

(3)FCP:

It’s Fiber channel protocol so, the data is transferred in the form of Frames

Each Frame size is 2 MB In Host side we are using HBA and Array side we are using Front

end ports It will supports Min 1Gbps & Max 8Gbps It will uses Fiber channel cables It has limited connectivity (180 Km’s only)

(4) IFCP:

It works with the concept of Gate way to Gate way communication

Gateway means it is nothing but a controller i.e., it allows outside network to inside and inside network to outside

Every connectivity should be in Fiber channel connectivity Cost should be very high and I/O performance is very high

1

Here Fiber Routers are used Router means connecting two different network

communications

(5)FCIP:

To reduce the cost we are using the FCIP instead of IFCP ,data transfer through internet

In FCIP we are using Ethernet technology Data transfer will be in the form of packets I/O performance is Slow Because Here we are using Ethernet

Based Protocol and cost should be low

(6)FCOE:

Fiber Channel Over Ethernet Actually, it is not yet implemented It support Minimum speed of 10 Gbps At Host Side, Instead of HBA’S we use CNA’S(Converge Network

Adapter) at the Host Here we need to change the entire H/W setup Instead of FC Switches we have to use FCOE Switches

FC LAYERS:

If we want to transfer the data from source to destination the data will move through FC layers, all the process will happens at backend. The process was completed within micro seconds

We have different layers; they are FC0, FC1, FC2, FC3, and FC4

FCO:

1

Check whether the physical connectivity established or not means HBA’S are inserted in PCI (Peripheral Component Interface) slots of servers(Hosts) are properly or not, frontend ports are working properly or not

FC1:

Encryption and Decryption are happened It will Encrypt the data with 8 bits while transform. Decryption the data with 10 bits while it reaches the destination For every 5 bits and 3 bits generates 1 parity bit each Parity means performs a XOR operations

FC2:

It manages frame sequencing and flow control Frames are transfer in sequence manner Each Frame size is 2Mb Frame contains Start of Frame, Frame Header & Data pay

Load(Includes Dest’n Id, Source Id,24 Bytes),CRC(CYCLIC REDUNDANCY CHECK,4 Bytes) and End of Frame(1 Byte)

Flow control is implemented by different classes Classes are of 5 types, They are:

Class-1: It has a dedicated connection and acknowledgement Class-2: It has a no dedicated connection and acknowledgement Class-3: It has a no dedicated connection and no acknowledgementClass-4: It has a fractional bandwidth with virtual circuitsClass-5: Multicasting with acknowledgement

FC3:

1

It provides common services required for advanced features, They are:

Stripping : It transmit single unit information to multiple links Group hunting: Mapping multiple ports into single port Multicasting: It delivers a single transmission into multiple ports Striping to multiply bandwidth using multiple N_ports in parallel to transmit a single

information unit across multiple links;

Hunt groups for more than one Port to respond to the same alias address

Multicast to deliver a single transmission to multiple destination ports.

FC4:

It uses to communicate with upper level protocol Network and channel protocols are specified as FC4 SCSI , IP, Single byte command code set mapping(SBCCS),link

encapsulation(FC_LE),Intelligent peripheral interface(IPI)

DISKS

In olden days we are using IDE(INTEGRATED DRIVE ELECTRONIC) Drives, IDE Drives introduced by Western Digital in 1986

Introduced a Drive controller being contained on the Drive itself Now a days we are using SATA Drives Disk performance should be maintained equal to fiber channel

Different types of disks:

1. ATA: Advanced Technology Attached2. SATA: Serial Advanced Technology Attached3. SCSI: Small Computer System Interface

1

4. SAS: Serial Attached SCSI5. FCD: Fiber Channel Disk6. SSD: Solid State Disk or Solid State Flash drives

(1)ATA:

Advanced Technology Attached Io requests to store into ATA Drives , First it will come to

processor then go to ATA Drives We have a different types of ATA Drives ATA1: Initially they introduced PIO (Processor I/O)[It is a method

of transferring data from a device to another device. Utilizing computer processor not DMA

ATA2:After some time not responding processor(Burden of Processor) come up with a concept called DMA(Direct memory access)[It is used to access the data directly without using CPU, It takes from RAM]

ATA3(1997):In ATA 3 uses SMART technology ,SMART(Self Monitoring Analysis Reporting Tool)I.e. If any bad sectors are occurring in hard disks, it will give alerts to user, user will closed that Bad Sectors

ATA4(1998): Packet interface is used , Data transfer rate is 33Mbps

ATA5:Approved in 2000,It supporting Data transfer Rate up to 66 Mbps

ATA6:Approved in 2001,It supporting Data transfer Rate up to 100 Mbps

ATA7:Approved in 2002,It supporting Data transfer Rate up to 133 Mbps

No Backward Compatibility

1

We can connect Max 2 devices per cable It has distance limitation up to 92 inches cable length It supports storage capacity 500GB Spindle Technology is used

(2)SATA:

Serial Advanced Technology Attached In SATA we have 2 models 1)SATA-1 2)SATA-2 The major difference b/w SATA & PATA is the point-to-point link

b/w Host and Drive Each device has a link to SATA to Host Port In SATA ,Max we can connect 2 devices this is the big

Disadvantage SATA supports SATA Drives only SATA used ATA commands In SATA-1:We used Bridge chip technique ,It will convert ATA

devices to SATA devices In SATA-2: We used NCQ(Native Command Query) technology

NCQ maintains the Log, it will shows where the data is presented in which sector, By using this, Directly goes to the location where the data is presented, that means I/O performance will increased or[By using this technique we can increase the accessing speed. Because it will eliminate some sectors and finally increase I/O performance, by using LOG File NCQ identify the data is stored in which Block]

It supports Max speed is 6Gbps It supports cable length 1 Meter It will support Max storage capacity 1TB

1

Spindle Technology is used

(3)SCSI:

Small Computer System Interface SCSI Classified into 2 types 1)Parallel SCSI 2) Serial SCSI In Parallel SCSI We can connect Max 16 Devices, Length Min 3

Meters and Max 12 Meters It supports Max speed is 1.5Gbps Across the Bus It supports Half Duplex It supports storage capacity 500GB Spindle Technology is used

(4)SAS:

Serial Attached SCSI In SAS we can connect Max 16,384 devices It has Backward compatibility is possible with SATA drives Backward compatibility means within enclose we can insert both

SAS and SATA Drives SAS use SCSI commands set In this we use point –to-point connection It supports Full Duplex It will supports Max 6Gbps speed It will support cable length 8 Meters It supports storage capacity 1TB Spindle Technology is used

(5)FCD:

1

In this model based on the disk configuration the speed will increased

It will support Max 8Gbps speed Performance is very High It has distance limitation up to 10 Kilometers Length In this , it will support different disk sizes like

73GB,146GB,300GB,360GB,400GB RPM speed is 15,000 Spindle Technology is used

(6)SSD:

There is No Spindle Technology, it supports Ray Technology Data should be Electronic Format So performance is High Not yet Released in Market

EMC

Array: By combining ‘N’ no of disks together and make huge storage space.

Array Models: In these different types of Array Models, They are CX, CX3 AND CX4

(1)CX (FISH SERIES): In this we have 3 types:

Cx-300 (SNAPPER)

1

Cx-500 (TARPON) Cx-700 (BARRACUDDA)

(2)CX3 (HAMMER SERIES): In this we have 4 types:

Cx-3 10 (TACK HAMMER) Cx-3 20 (JACK HAMMER) Cx-3 40 (SLEDGE HAMMER) CX-3 80(HAMMER HEAD)

(3)CX4: (This series don’t have naming conversion) in this we have 4 types:

CX4-120(NAUTILUS) Cx4-240(IRON CLAD) Cx4-480(TRIDENT) Cx4-960(DREAD NAUGHT)

The above all are EMC Array Models The above all FC Arrays, These are Mid Range Arrays

CX-300 ARCHITECTURE:

In this we have two types of Enclosures DPE (disk processor enclosure) , DAE (disk array enclosure)

In DPE from Front End side we have disks, from Back End side we have controllers, And In DAE from Front End Side we have disks, from Back End Side we have Primary and Extension ports

We have two controllers SP-A and SP-B, array is totally controlled by these controllers, the two controllers SPA & SPB

1

are always synchronized, the communication happens b/w controllers by using CMI (clarion message interface)

In CX-300 we have 1 Back End Port per controller and 2 Front end ports per controller,

Back End Ports is used to communicate b/w the enclosures (bus connection), Front End Ports is used to communicate with the host.

Back End Ports are connected to the enclosures with copper FC cable (1-2 meters length), array BE cabling is implemented by orbited loop only and the connectivity is called link control cards (LCC). And Optical FC Cables is used for Front End Ports

We can manage the array by using NAVISPHERE Manager Software.

LAN Port is used to assign IP address for multiple Hosts or We can access the array throughout the network[Remotely]

COM1 Port is used for find out the trouble shoot the issues, whenever the controller gone bad

In an Enclosure, for every 5 disks there is a light indication[LED] CX-300 Will support ATA,FC

CX-500 ARCHITECTURE:

In this we have two enclosures DPE (Disk Processor Enclosure) , DAE (Disk Array Enclosure)

In DPE from Front End side we have disks, from Back End side we have controllers, And In DAE from Front End Side we have disks, from Back End Side we have Primary and Extension ports

We have two controllers SP-A and SP-B, array is totally controlled by the controllers, the two controllers SPA & SPB are

1

always synchronized, the communication happens b/w controllers by using CMI (clarion message interface)

Cx-500 has 2 Back End Ports per controller and 2 Front End Ports per controller.

Back End Ports is used to communicate b/w the enclosures (bus connection), Front End Ports is used to communicate with the host.

Back End Ports are connected to the enclosures with copper FC cable (1-2 meters length), array BE cabling is implemented by orbited loop only and the connectivity is called link control cards (LCC).

We can manage the array by using NAVISPHERE Manager Software.

LAN Port is used to assign IP address for multiple Hosts or We ca access the array throughout the Network[Remotely]

COM1 Port is used for find out the trouble shoot the issues, whenever the controller gone bad

In an Enclosure, for every 5 disks there is a light indication[LED] CX-500 Will support SATA,FC

CX-700 ARCHITECTURE:

In this we have two enclosures SPE (Storage Processor Enclosure), DAE (Disk Array Enclosure).

In this we use SPE instead of DPE, in this we have 3 Jumbo fan packs from Front End Side, and Back End Side we have controllers(Top and Bottom) , In DAE from Front End Side we have disks, from Back End Side we have Primary and Extension ports

We have 4 Back End Ports and 4 Front End Ports per controller

1

CX-700 will support FC, SATA, SAS drives

Types of Enclosures: we have three types of enclosures Katana(FC 2 Gbps)drives Klondike (ATA,SATA)drives Stiletto (FC 4Gbps,SAS)drives

DIFFERENCES BETWEEN CX-300,CX-500,CX-700

Types CX-300 CX-500 CX-700

Names SNAPPER TARPON BARACCUDA

No. of Enclosures 4 8 16

No. of Disks 60 120 240

Types of Enclosures Katana/Klondike Katana/Klondike Stiletto/Klondike

BE Ports/FE Ports 1/2 per controller 2/2 per controller 4/4 per controller

Type of disk support FC/ATA FC/SATA FC,SAS/SATA

Max No. of Hosts 64 128 256

Max No. of LUNS 512 1024 2048

Max LUN size 512 GB 1 TB 2 TB

Max Cache size 2 GB 4 GB 8 GB

Max Array size 27 TB 59 TB 119 TB

DIFFERENCES BETWEEN CX3-10,CX3-20,CX3-40 & CX3-80

Types CX3-10 CX3-20 CX3-40 CX3-80Names TACK

HAMMERJACK HAMMER

SLEDGE HAMMER

HAMMER HEAD

No. of Enclosures

4 8 16 32

No. of Disks 60 120 240 480BE Ports 2 2 8 8FE Ports 4 12 8 8Max Host connections per Array

128 256 256 512

Max No. of LUNS

512 1024 2048 2048

Max LUN size 512 GB 1 TB 2 TB 2 TB

1

Max Cache size

2 GB 4 GB 8 GB 16 GB

Max Array size

60 TB 114 TB 234 TB 474 TB

DIFFERENCES BETWEEN CX4-120,CX4-240,CX4-480 & CX4-960

Types CX4-120 CX4-240 CX4-480 CX4-960

No. of Enclosures

8 16 32 64

No. of Disks 120 240 480 960BE Ports(FC/iSCSI)

2/4 4/4 4/4 8/4

FE Ports(FC/iSCSI)

4/4 4/4 8/4 8/4

Max No. of LUNS

1024 2048 4096 8192

Max No. of Meta LUNs

512 512 1024 2048

No. of Raid Groups

60 120 240 480

Max cache size 6 GB 8 GB 16 GB 32 GBMax Array size 120 TB 231 TB 939 TB 1899 TBCPU clock rate 1.2 GHZ 1.6 GHZ 2.2 GHZ 2.33 GHZ

FLARE (FIBER LOGIC ARRAY RUNTIME ENVIRNOMENT):

It is similar to normal Operating system but it’s an Embedded XP operating system

The flare will be installed on 1st Enclosure of 1st Five drives These drives are called “vault” drives. They are numbered as

0,1,2,3,4 While installing the flare it will take 6.4gb from each drive to

install the flare and remaining space is used for user storage

1

But EMC will not recommend to use the free space of the vault drives

Flare will install on 4 drives and 5th drive is used of vault area The 0,2 will take care by SP-A and 1,3 will take care by SP-B , 4

drive is for log’s (vault area) In vault drive along with flare ,it contains

1. Vault Area 2. PSM (Persistent Storage Management) LUN 3. Flare LUN

(1)Vault Area:

It is the save area for write cache in case of SP failure or power failure of the CLARiiON

(2)PSM (Persistent Storage Management) LUN:

It maintains the total configuration of the CLARiiON Internally its support RAID Type 3 These are located on first 3 drives in vault drives. It occupies 3

GB space from the 3 drives

(3)Flare LUN:

It contains flare code which is running on CLARiiON

NOTE:

Without using the 5th disk we can install the FLARE but we cannot enable the write cache

How to Install FLARE:

1

To Install the FLARE there is a separate process called ICA (Image Copy Application)

To Install FLARE we use a separate mission called ICA server 1st we need to insert vault drives (0-4) in ICA server and then

download the FLARE IMAGES(bundles) it contains flare code Then we have to install that code into vault drives After installation remove those drives from ICA server and

relocate them in same order in first Enclosure of the array To Bring up the Array it will take 1 hour time FLARE is mediator b/w Array and user

Command for FLARE Installation in ICA:

C:\flare install “Path Of FLARE IMAGES” ,SPA IP, subnet mask, gate way , SPB IP, subnet mask ,gate way;

SPA subnet mask is 250.0.0.0 & SPB subnet mask is 255.255.0.0

NDU (NON DISRUPTIVE UPGRADE) Process:

If we want to install any Array Based S/W on the controller, we need to follow the process called NDU Process

The Extension of NDU Process is “.pbu”Array Based S/W’S:

Navisphere Manager Access Logic’s Snap View Mirror View San Copy Flare patches

1

SPA=10 LUNS & SPB=10 LUNS In internet explorer whatever the IP address given then the NDU

process start on that controller first. Ex: http:// IP address of SPB

We start NDU process on SPB On that time the controller SPB down then the 10 LUNs are

“Trace Pass” to controller SPA Then the SPA take care about the total Array After completion of NDU process on SPB, it will send the

Acknowledgement to SPA, Yes I am safe. Then the 20 LUNs are “Trace Pass” to SPB and NDU process starts on SPA controller

After completion of NDU process on SPA the LUNs that are owned by SPA will “Fail Back” to SPA

The software that doesn’t impact immediately until unless you have to hit the “Commit” option.

NOTE:

TRACE PASS: We can pass the LUNs to one controller to another controller manually is called “Trace Pass”

NDU process will takes place on single controller at a time, from which controller you send a request to perform NDU on that controller only, it will perform first, after completing the upgrading on that controller next it will perform on another controller

SPA SPB

1

CLI (Command Line Interface):

It is similar to Navi Agent but this is Command Line Interface It is a command prompt It has two types. They are (a) Navi CLI (b) Navi secure CLI

(a) Navi CLI: In this while executing the commands we have to login with User name and Password only ONCE

(b) Navi secures CLI: In this while executing the commands we have to login for each and every command with User name and Password

NAVI AGENT:

It is a host based s/w, we need to install on the server Every Host have World Wide Number[WWN] The AGENT will communicate with the Array controllers Whenever we have a Navi Agent at host side, it will automatically

register the host information in Navi Sphere Manager Whenever we install a host based s/w (Navi Agent) on host

side ,first it will communicate to controller and automatically take the information of new host is registered in Navi Sphere Manager

NOTE:

Whenever we don’t have a Navi Agent at host side, we can do manually the register the host information in Navi Sphere Manager

How to register the new host information in Navi sphere manager by manually:

Right click on the array then go to connectivity status

1

In that window select group edit button, it will display the group edit initial window

In that window Left-Hand pane have a available hosts are there and Right-Hand pane have a selected hosts are their

And select Array commpath as well as Failover Mode Click the New Host button and Give the Hostname: ……….

IP address: ……… then select <ok> Button

LAYERED APPLICATIONS: These are used for Local and Remote Replications.

We have different types of Layered applications are there. They are

1) Snap view (a) Snap shot (b) Snap clone 2) Mirror view (a) M/V Synchronous (b) M/V Asynchronous 3) San copy (a) Incremental copy (b) Full copy

(1) SNAP VIEW: We used for Local replication or backup purpose

By using snap view we can create “Snap Shot” & “Snap Clones”

(a)SNAP SHOT: It is point-in-time-copy of source LUN within the Array. Snap shot will works with a concept of copy on 1st write.

We can create 8-snap shots. Snap shot takes memory from RLP

(b)SNAP CLONE: It is an exact copy of your “source LUN” and will maintain the link with your “source LUN” always. After creating “Snap clone” whatever the data we write to your source LUN that will automatically replicated to your “Clone LUN”. (2) MIRROR VIEW: It is used for Remote Replications or Remote

1

backups. By using Mirror view we can create “Mirror clones”. In Mirror view we have two types:

Mirror view Synchronous Mirror view Asynchronous

i)It supports limited distance. It supports 180 KM only.

ii)Performance is “slow” because when the I/O’s coming from host to primary array then it replicates to secondary array, then after secondary array sends “Ack” to primary array from here to host

i)It is unlimited distance

ii)Performance is “High” because when the I/O’s coming from host to primary Array then Primary Array sends “Ack” to host then it replicates to secondary Array

(3)SAN COPY: This is used for Migration purpose. It may be LUN Migration OR ARRAY Migration

We have to Migrate the array from CX-700 to CX4-120 FCIP & IFCP Protocols are used(For transferring the data one place

to another place) This is the tool which supports to multiple vendor like HP, IBM,

HITTACHI….. Here we have two types of copies: (a) Incremental copy (b) Full

copy

(a)INCREMENTAL COPY: In this the total data in the array will be moved to another array. After that it will maintains a link between the two arrays

1

(b) FULL COPY: In this the total data in the array will be copied to another array. After that there is NO link between the two arrays

HBA (Host Bus Adapter):

It is a hardware component which is able to communicate with the array.

Without HBAs we can’t communicate with the array HBA vendors are (i) Q-Logic (ii)Emulex We need to insert the HBAs in “PCI” slots of servers This HBA’S also have Management software

For Q-Logic San surfer and For EmulexHBA Any ware

If you want to see the WWN in HBA Management software i.e., startprogramsQ-LogicSan surfer startprogramsEmulexHBA Any ware

In the Management software of HBA we can see the WWN’S, BIOS version, HBA BIOS version, HBA Driver version

HBA’S are needed to connect the server(Host) to the storage By using the HBA’S we can see the storage Every HBA has WWN & Every HBA has WWPN

POWERPATH:

Power path is a Multipathing software which is provided by EMC The major functionality of the power path is Load balancing and

path management

1

It is a host based software CLARiiON supports on active-passive architecture If power path is there then it will equally distributed the I/O’S

across all the paths

(1)LOAD BALANCING: Whenever we have the two ports HBA’s then only the Load balancing happen. Distributing the I/O’s are equally shared to all the active paths two types of load balancing are there: (a) Static Load balancing (b) Dynamic load balancing

(a)STATIC LOAD BALANCING: It is performed by manually by the Administrator. Luns are balanced across SP’s by changing the ownership of those Luns.

(b)DYNAMIC LOAD BALANCING: Power path spread I/O work load across all available paths

This will provide better utilization across HBA’s but doesn’t necessarily balance workload across storage processors

(2) PATH MANAGEMENT: It provides path failover i.e., whenever the working path is fail, it will sends I/O’s from another path.

NOTE:

FAILOVER : If any power path is gone bad, I/O’s will move to other path called Failover (Automatically).

FAIL BACK : Once power path is resolved, I/O’s will come back i.e., Fail Back.

POWERPATH POLOCIES:

1

(i)Claropity: I/O requests are balanced across multiple paths based on composition of reads, writes and user assignable priority.

(ii)LEAST PATH I/O’s: I/O requests are assigned to the path with the fewest number of requests in the queue

(iii)LEAST BLOCK I/O’s: I/O request are assigned to the path with the fewest total blocks in the queue

(iv)ROUND ROBIN: I/O requests are distributed to each available path in rotation manner

(v)REQUEST: No load balancing but it will maintain path management

(vi) BASIC FAILOVER: Provides SP Failover for single HBA system

NOTE:

Policy may be sent on a LUN by LUN basis. CLAROPITY is the default and appropriate for nearly all

environments

NOTE:

Array comm. path: The Array comm. Path option provides a communication path to

a storage processor, even when there are no LUNS configure for the Host

FAILOVER MODE: The FAILOVER MODE option affects how FAILOVER is initiated Array initiated or Host initiated Set to 1, if Host is running Power path

1

Set to 0, if no FAILOVER s/w is in use

Cache: In Hp we have two types of caches, they are 1. Control cache

2. Data cache

(1)Control cache: when the I/O’s are coming from host, first the I/O’s will come to controllers after that controllers to disks

Whatever the data coming from controllers to disks that data will take care by “control cache”.

(2)Data cache: Whatever data coming from host to controllers that data will take care by data cache.

HP

INTRODUCTION:

It is a SAN based storage It is high performance and virtual raid storage HP supports Vraid-0, Vraid-1,Vraid-5,Vraid-6 In HP arrays, space utilization is very high because these are

“Virtual Raid Storage” It supports Local and Remote Replications HP supports Max Array Size is 324TB In HP we have two HSV controllers, they don’t have

identification HP we have two loop switches, it provides communication b/w

enclosures

1

HP arrays are managed by a s/w called Command View, it’s a GUI tool

If we want Access the Arrays Remotely, then we need to give the “IP Address” of the particular host which have the CV and along with Port no http:// IP Address: 2372

HP is EVA (Enterprise Virtual Array), it is also a Mid Rang Array.

In HP we have different types of Array Models:

ARRAY MODEL CONTROLLER MODELEVA 4400(LE) HSV 300EVA 6400(DEXL LOW) HSV 400EVA 8400(DEXL HIGH) HSV 450

Architecture of EVA 4400:

This is array based management S/W, it contains a management module at Back End

We have to assign the IP address to that module for identification

We have to install the CV into that module , CV will support max 32 arrays

In this we have two embedded switches, it supports 8Gbps speed

Each switch contains 12 ports In this we have “2” HSV 300 controllers In this we have 1 Back End Ports & 2 Front End Ports per

controller Array will support 1-8 Enclosures

1

Each Enclosure contains 0-11 disks It will support Max 96 Disks Cache size is 1 GB per controller

Architecture of EVA 6400:

This is a server based management S/W, we need to have a separate server to manage these array

We need to install the CV in that server, it requires Win2k3 operating system

If we want to access the array from our desktop, we need to give the IP address and port number

In this we have “2” HSV 400 controllers In this We have 2 Back End Ports & 4 Front End Ports per

controller Array will support 2-18 Enclosures Each Enclosure contains 0-11 Disks It will support Max 216 Disks Cache size is 2 GB per controller

Architecture of EVA 8400:

This is a server based management S/W, we need to have a separate server to manage these array

We need to install the CV in that server, it requires Win2k3 operating system

If we want to access the array from our desktop, we need to give the IP address and port number

1

In this we have “2” HSV450 controllers In this we have 3 Back End Ports & 4 Front End Ports per

controller Array will support 3-27 Enclosures Each Enclosure contains 0-11 Disks It will support Max 324 Disks Cache size is 4GB r 8GB per controller

HP ENCLOSURES:

We have different types of Enclosures are there. They are(1)MOONGAZER (2) CALLISTO (3) CAMARO

(1) MOONGAZER:

It supports “FC” and “FATA” drives We insert the drives Vertically in this Enclosure This model is used in EVA,LE & DEXL It supports Max 14 drives initially FATA-Fiber Attached Technology Adapted

(2) CALLISTO:

It supports “FC”,”FATA” & “SSD” drives We insert the drives Horizontally in this Enclosure This model is used in EVA 4400,EVA 6400 & EVA 8400 It supports Max 12 drives and it supports Max 4 GB

1

(3) CAMARO:

It supports “SATA” & “SAS” drives We can arrange the drives in two ways

(i)Large Form Factor (LFF) 12 drives only(ii)Small Form Factor (SFF) 25 drives only

This model is used in WIZER It supports Max 8 GB

DISK GROUP:

It requires Min 8 drives In HP we will insert the drives Horizontally 0-11 drives we can insert per one Enclosure If we want to use the drives, we have to create the Disk group

otherwise there are called Ungrouped drives.

NOTE:

LE-Lower End HP Arrays will support the following Raid Levels VRaid-0, VRaid-

1, VRaid-5 & VRaid-6 VRaid-6 will support the Firm Ware 9500 onwards HSV-HP Storage Virtualization

TABEL FOR LE, DEXL LOW & DEXL HIGH:

TYPE EVA 4400 ( LE)

EVA 6400 (DEXL LOW)

EVA 8400 (DEXL HIGH)

Max No. of Enclosures

8 18 27

Max No. of Disks 96 216 324Enclosure Model Num & Type

6412 (CALLISTO) 6412 (CALLISTO) 6412 (CALLISTO)

BE Ports/controller 1 2 3

1

FE Ports/controller 2 4 4Mirror port controllerController Model Num

HSV 300 HSV 400 HSV 450

Max No. of LUNs 1024 2048 2048Max LUN size 32 TB 32 TB 32 TBCache Memory/controller

1 GB 2 GB 4 GB or 8 GB

Policy Memory/controller

1 GB 2 GB 3 GB

Array Type Array Based Management

Server Based Management

Server Based Management

RSS (Redundant Storage Set):

It means sub grouping of physical drives for failure separation at the Redundancy level

Whenever we are creating volume it will allocates space of RSS The RSS size is 8 MB

LAYERED APPLICATIONS:

Here we have two types of layered applications are there, they are:(1)Business Copy (BC) Used for Local Replications (a) Snap shot (b) snap clone (c) Mirror clone(2)Continuous Access (CA) Used for Remote Replications (a) Mirror view synchronous (b) Mirror view Asynchronous

(a)SNAP SHOT:

Point-in-Time-copy In HP snap shot uses “Copy Before Write Technique”

1

After creating the snap shot the data copy into the destination when the next I/O’s are happened in the source

Whenever we creating the snap shot the space is allocated from Disk group

Here we have two different snap shots are there: (i)Demand Allocated snap shot (ii)Fully Allocated snap shot

(i)DEMAND ALLOCATED SNAP SHOT: In this after creating the snap shot, the size is “0”, while the data copying into that snap shot, the size will increase.

(ii)FULLY ALLOCATED SNAP SHOT: In this the snap shot will creates with the Max size of source LUN size. So there is no risk of over commitments. In this the data also copied into snap shot.

(b)SNAP CLONE:

Exact copy of source LUN The snap clone whatever the data we have in the source LUN

that data will Replicate in the destination or clone LUN After creating the snap clone, the clone LUN act as independent

LUN because there is NO link between source LUN and clone LUN, We can present this clone LUN to the Host.

LOAD BALANCING POLICIES IN HP:

Load balancing is used to distribute I/O request across multiple paths to logical unit for optional usage of paths and improve I/O performance.

1

The DSM(Device Specific Module) currently supports the following Load balancing policies

(1)ROUND ROBBIN (RR): This policy assigns equal priority to all the paths available for a device and routes I/O request to each path in order

(2)SHORT QUEUE SERVICE TIME (SQST): This policy selects the path based on service time for the outstanding I/O requests and chooses the path with the least outstanding I/O service time for the incoming I/O requests. This is “Default policy” set by DSM.

(3) SHOTEST QUEUE REQUEST (SQR): This policy selects the path with the least outstanding I/O request

(4)SHORTEST QUEUE BYTES (SQB): This policy selects the path with least outstanding I/O bytes

(5)NO LOAD BALANCE (NLB): This policy doesn’t not implement any I/O load balancing, However the DSM selects a path to service all the I/O request it preferred paths is not set. All the devices that are part of the logical unit group have the same load balancing policies.

1

SWITCHS

Switch:- Switch is a small hardware device that joins multiple computers together within one local area network (LAN) .Brocade Switch:-Hardware architecture:ASIC: Application Specific Integrated Circuits

ASIC generations:-1. Switch 512mbps 2 port2. Loom 1gbps 4ports3. Bloom 2gbps Bloom 1(Low level companies)

Bloom 2 (Midrange, High level companies) Bloom 1:- 3200 (8 ports), 3800(16 ports) Bloom 2:- 3900 (32 ports), 12000(128 ports)4. Golden eye 1, 2 (4gbps)5. Contour 1, 2

FPROM

CPU

1

Switch status commands: - Brocade switch in the following waysTemp show: - it shows condition of the temperatures of the devices.Fan show: - it shows the working condition of the fans.Ps show: - it shows the power supply status.Default IP for Brocade switch is 10.77.77.77We can access Brocade switch in the following ways

telnetSshHttpSecure TelnetSmi-SSnmpFabric manageBrocade API

Brocade Switch level roles:UserAdminFactoryRoot

Switch levels: 1. License show2. Switch Disable3. Configure

If you want to configure your switch, you need to disable the switch.Ports level:Port cfgshow: show all switch port summary.Port show number: Shows individual port number

Different Types Of Ports: We have different type of ports,1. N-port 2.NL-port 3.F-port 4.FL-port

1

5. ISL 6.TE –port 7.E-port.

1. N-port: N-port means Node port. N-port attaches an F-port in a point – point protocol.

2. NL-port: NL-Port means Node Loop port. Node will be connected in the loop. It transmit the load is connected to the receive load

3. F-port :F-Port means Fabric port. The switches have F-ports, it is connect a N-port on one-one basis for connect a server to a storage nodes in a fabric sequence.

4.FL-port :FL-Port means Fabric Loop. It is a fabric loop port is used for connection in san

5.ISL :ISL means Inter Switch Link. It is used to connect two different switches by using E-port (Expansion ports).

6.TE-ports : TE-Port means Trunk Enable Port. We establish trunk enable between two switches it will broadcast the speed across two switches.EX: we have two switches both working with 2gbps speed and we establish trunk enable between these two switches it will broad cast the speed as 4gbps speed this is the use of the trunk enable.

7. E-Port :E-Port means Expansion port . It can be used to establish ISL between two switches.

F-logi:-When you are connecting physically from nodes to Switch, the node will send its information (64 bit address) in a frame to switch that is called F-logi.

1

P-logi:- Switch will convert the 64bit address in to 24 bit and save in to the name server, after that Switch will send 24bit address to node, that is called P-logi.

Zoning

Definition: - Limiting the scope of the initiator is called Zoning. it is restricting only storage sub-system ports not Luns, (Giving the restricting between the host and array).

There are two types of Zones 1. Hard zone

2. Soft zone

1. Hard zone: - By using switch port number and domain id we can create “Hard Zone”, if any one of the switch port is gone bad we have to change the zone configuration.

It is difficult to manage in large san’s, but provides more security.

Hard Zone maps devices physically to a port. In this if a switch gone bad ,we can’t simply plug a cable in

another port’s It provides more security.

2. Soft Zone:- By using array wwn and host wwpn’s we can create “Soft Zone”.

If any one of the switch port is gone bad, here no need to change the zone configuration, just unplug the cable from failure port and plug it in new port.

1

If any HBA fails we have to change the zone configuration. Soft Zone maps the devices logically. It is more flexible and easy to implement in large complex San’s.

We can implement Lun security in 3 places.

HBA Selecting lun presentationSwitch ZoningStorage Lun masking

Process to configure the zone:-

1. Identify each device and its wwns or ports where the device is connected.

2. Assign a meaning full alias name for each device r port.3. Create zones and include desired devices in them.4. Create configuration and include required zones in it.5. Save and enable the configuration.

1

VM WARE

VMware is a company that was found with a concept of virtualization.

1. By using virtualization we can manage multiple OS in one system.

2. It will decrease the physical burden on the hardware, like creating virtual components for our hardware

3. In this virtualization we have two servers. 1. GSX server (group server virtualization)2. ESX server (enterprise server virtualization).

GSX server:-

1. This server will support windows Operating System, windows is the traditional for GSX server.

2. Hard ware is located in the bottom of the GSX server.3. On the top of the hardware traditional OS is stored.4. On the top of the OS we have to install the VM workstation or

VM stack.5. On the top of the workstation we can create Virtual missions or

guest OS.6. Whatever the Os you require, you can load that in the VM’s.7. VM’s will support all Operating Systems like windows, Linux,

Solaris.

1

8. While creating the VM’s it will create a profile with all hardware components.

9. Once we load the OS in VM, the VM will act as a normal OS.

ESX server:-

1. This server will support Linux OS.2. This OS will act as a work station in the ESX server.3. On the top of the OS we can create VM’s.4. To present the Lun’s to the VM’s, we have two file systems.

1. RDM (raw device mapper)2. VMFS (virtual mission file system)

5. ESX server will support max 2TB lun, the size will depends up on the Block size.

RDM: - by using RDM we can present the total lun to only one VM, we can’t share the Luns to all VM’s.

VMFS: - by using VMFS we can present the Luns to all VM’s, we can share the Luns.

We have two software’s to connect ESX servers, they are 1. VMIC (virtual mission infrastructure client) 2. VMIS (VM Infrastructure center/server)

By using VMIC we can manage one ESX server along with VM’s in one pc.

By using VMFS we can manage multiple ESX servers in one pc.

1

V-motion: - It is software; by using this we can do remote replications from one ESX to another ESX server and one VM to another VM.