dplug

1. Всегда скопируй dplug файд иначе он исчезнет после того как ты его загрузишь:copy bootflash:<dplug_name> bootflash:dplug_temp
2. Загрузи dplug: load bootflash:dplug_temp
3. Что бы посмотреть PID процесса, например fcns: ps -ef | grep fcns
4. Чтобы убить и перезапустить процесс: kill -9 <PID>

Чтобы правильно выбрать dplug, надо проверить полностью имя image который установлен на свиче: 

trabzon# show ver | grep image

  kickstart image file is: bootflash:/m9500-sf2ek9-kickstart-mz.4.2.9.bin

  system image file is:    bootflash:/m9500-sf2ek9-mz.4.2.9.bin

trabzon# show ver | grep MDS

  cisco MDS 9506 (6 Slot) Chassis ("Supervisor/Fabric-2")

Нам нужен dplug который для версии 4.2.9 и имеет sf2ek9 в названии. Кроме того он должен быть для платформы 9500.

Чтобы найти нужный файл:

1. telnet people.cisco.com
2. Перейти в нужную директорию в зависимости от ОС версии:

   Release	Location
   3.x	/auto/androme/REL_3_1_y_z
   3.2	/auto/andau/REL_3_2_y_z
   3.3-4-5	/auto/andr33/REL_3_3_y_z
   4.x	/auto/andankara/REL_4_x_y_z
   5.0.x	/auto/andbogota/REL_5_0_x_y_z
   5.1.x	/auto/andcairo/REL_5_1_x_y_z
   5.2.x	/auto/anddelhi/REL_5_2_x_y_z
   6.0.x	/auto/andedinburgh/REL_6_0_x_y_z
   6.1.x	/auto/andedinburgh/REL_6_0_x_y_z
   6.2.x	/auto/andfreetown/REL_6_2_x_y_z

   
   
3. cd build/images/gdb
4. ls *dplug*
5. copy the dplug image to ftp.cisco.com

DCNM overview

SCSI LUN discovery

SCSI LUN discovery feature can be used to find all targets (Disks and Tapes) and their LUNs(Logical Unit NUmbers) connected the local switch and/or remote switches.

SCSI LUN discovery feature uses -FFFFFD FCID address as a source to perform SCSI INQUIRY, REPORT LUNS, READ CAPACITY commands on SCSI devices.

mds9124(config)# show scsi-target disk 
--------------------------------------------------------------------------------
VSAN    FCID        PWWN                      VENDOR    MODEL             REV
--------------------------------------------------------------------------------
mds9124(config)# discover scsi-target local os all
discovery started
mds9124(config)# discover scsi-target remote os all
discovery started
mds9124(config)# show scsi-target status
discovery completed
mds9124(config)# show scsi-target disk
--------------------------------------------------------------------------------
VSAN    FCID        PWWN                      VENDOR    MODEL             REV
--------------------------------------------------------------------------------
10      0x250000    50:0a:09:82:99:9b:48:e0   NETAPP    LUN               811a

DPVM = Dynamic VSAN Port Membership

If task states that server HBAs can be moved from one MDS to another - time to use DPVM!

Step1: create VSAN an assign server's port to it.
mds9124(config)# vsan database
mds9124(config-vsan-db)# vsan 20 name BLUE
mds9124(config-vsan-db)# vsan 20 interface fc1/3
Traffic on fc1/3 may be impacted. Do you want to continue? (y/n) [n] y
mds9124(config-vsan-db)# exit

Step2: Create DPVM database for the HBA(s):

mds9124(config)# dpvm enable ! tab does not work on this command. You should type it manually

mds9124(config)# dpvm database

mds9124(config-dpvm-db)# device-alias Initiator vsan 20

Error: Device-alias dpvm  member unsupported in device-alias basic mode

!need to have enchanced mode for device-alias

mds9124(config)# device-alias mode enhanced

mds9124(config)# device-alias commit

mds9124(config-dpvm-db)# device-alias Initiator vsan 20

Successful. Commit should follow for command to take effect.

mds9124(config)# dpvm commit

Commit operation initiated. Check dpvm session status

Step3: activate DPVM database with a force parameter (without force commit might fail)

mds9124(config)# dpvm activate force

Successful. Commit should follow for command to take effect.

mds9124(config)# dpvm commit

Commit operation initiated. Check dpvm session status

Step4: verify the results

mds9124(config)# show dpvm status

DB is activated successfully, auto-learn is off, distribution is enabled,

Duplicated pwwn will be  Rejected.

mds9124(config)# show dpvm database

device-alias Initiator [21:00:00:e0:8b:0a:fc:6d] vsan 20

device-alias TestFcalias [50:0a:09:82:99:9b:48:e0] vsan 20

[Total 2 entries]

!In order to automatically copy DPVM configuraiton, on other switches you should do:
1. dpvm enable - on other switches
2. dpvm commit - on switch which has the configuration

Basic interface and VSAN configuation

mds9124(config)# interface fc1/5
mds9124(config-if)# switchport description Connection to Lancelot fc1/12
mds9124(config-if)# switchport speed 4000
mds9124(config-if)# switchport mode E

mds9124(config)# vsan database

mds9124(config-vsan-db)# vsan 10 name RED

mds9124(config-vsan-db)# vsan 10 interface fc1/3

Traffic on fc1/3 may be impacted. Do you want to continue? (y/n) [n] y

mds9124(config)# show vsan membership

vsan 1 interfaces:

    fc1/1             fc1/2             fc1/4             fc1/6

    fc1/8             fc1/9             fc1/10            fc1/11

    fc1/12            fc1/13            fc1/14            fc1/15

    fc1/16            fc1/17            fc1/18            fc1/19

    fc1/20            fc1/21            fc1/22            fc1/23

    fc1/24

vsan 10 interfaces:

    fc1/3           

mds9124(config)# show vsan

vsan 1 information

         name:VSAN0001  state:active

         interoperability mode:default

         loadbalancing:src-id/dst-id/oxid

         operational state:up

vsan 10 information

         name:RED  state:active

         interoperability mode:default

         loadbalancing:src-id/dst-id/oxid

         operational state:up

! VSAN configuration should me manually done on all switches
! VSAN should ALWAYS be configured in the VSAN alowed trunk list otherwise all configured VSANs wil use the trunk.

Fibre Channel Addressing - FCIDs

MDS Fabric Switch assigns a VSAN-wide unique FCID to each device (HBA Port or Disk Drive) during the FLOGI. Other vendor switches assign a banric-wide unique FCID to each device.

The 24-bit FCID consists of three 8-bit fields:
1st byte - Domain ID of a switch the device is connected to.
2nd byte - Area ID - for each N_Port and NL_Port, the switch will assign area field started at 0x00 and incrementing by one for each consecutive ogin
3rd byte - Device ID 

Well-known FCID addresses:

• 0xFFFFFA - Management Server
• 0xFFFFFB - Time Server
• 0xFFFFFC - Name Server/Directory Server (show fcns database)
• 0xFFFFFD - Fabric Controller (Used for communication between switches)
• 0xFFFFFE - Fabric F_Port (Used by N_Ports for Fabric Login)

Fibre Channel Addressing - WWNs

WWN (Worldwide Name) is a unique 64-bit value that is assigned to a device at the factory.

Each N_Port and NL_Port (HBA port or Starage Disk) has one nWWN and one or more pWWNs (port WWN):

canterbury# sho flogi database
---------------------------------------------------------------------------
INTERFACE  VSAN    FCID            PORT NAME               NODE NAME
---------------------------------------------------------------------------
fc1/5       1     0xef0000  50:0a:09:81:99:9b:48:e0  50:0a:09:80:89:9b:48:e0     [Netapp]
fc1/12     1     0xef0003  50:05:07:63:0e:00:08:f3  50:05:07:63:0e:fe:08:f3        [Panda]
fc1/16     1     0xef0004  10:00:00:00:c9:32:27:85  20:00:00:00:c9:32:27:85

Total number of flogi = 3.

Each F_Port (physical port on the switch) has one fWWWN (fabric WWN)

mds9506# show interface fc 1/1

fc1/1 is down (SFP not present)

    Hardware is Fibre Channel

    Port WWN is 20:01:00:05:30:00:9b:9d

Each Fabric Switch has one sWWN (switch WWN)

mds9506# show wwn switch

Switch WWN is 20:00:00:05:30:00:9b:9d

Each VSAN has its own VSAN sWWN.

mds9506# show fcdomain vsan 1

The local switch is the Principal Switch.

Local switch run time information:

        State: Stable

        Local switch WWN:    20:01:00:05:30:00:9b:9e

        Running fabric name: 20:01:00:05:30:00:9b:9e

Static FC routes

Configures a static route in VSAN 5 for specific FCID 0x501010 to route traffic on the interface fc1/2. The neighbor switch connected on the interface fc1/2 has domain ID of 12:

fcroute 0x50101010 interface fc1/2 domain 12 vsan 2

Configures a static route in VSAN2 for any FCID beginning with 0x50xxxx to route traffic on the interface fc1/2. The neighbor switch connected on interface fc1/2 has domain ID of 12:

fcroute 0x500000 0xff0000 interface fc1/2 domain 12 vsan 2

Configurares a static route in VSAN 2 for specific FCID 0x501010 to route traffic on the interface fc1/2. The traffic should flow through a non-directly connected switch with domain ID of 13:

fcroute 0x501010 interface fc1/2 domain 13 remote vsan 2

FSPF = Fabric Shortest Path First

FSPF is a path selection protocol. IT is enabled by default on all storage switches and doesn't have to be configured.

1. Fibre Channel routing is based only on Domain ID (first byte in the 3-byte FCID address assigned by and MDS switch to an end device)
2. FSPF runs only on E and TE ports and it provides a loop free topology
3. FSPF runs on a per VSAN basis
4. FSPF routes are unidirectional. It is possible to have different paths in each direction of the fibre channel traffic
5. uses the Dijkstra's algorithm.
6. FSPF treats a port channel as a single logical link. If one of the members fails in a port channel FSPF is not aware of this change

Default FSPF cost:

1 Gig link is 1000

2 Gig link is 500

4 Gig link is 250

8 Gig link is 125

10 Gig link is 100

We could change this value:

(config)# interface fc 1/2

(config-if)# fspf const 20 vsan 2

mds9506# show fspf internal route vsan 1
FSPF Unicast Routes
---------------------------
 VSAN Number  Dest Domain   Route Cost    Next hops
-----------------------------------------------
           1    0xc9(201)          250        fc1/8
                                             fc1/12


bdc-mds9506-2# show fspf interface
FSPF interface fc1/8 in VSAN 1
FSPF routing administrative state is active
Interface cost is 250
Timer intervals configured, Hello 20 s, Dead 80 s, Retransmit 5 s
FSPF State is FULL
Neighbor Domain Id is 0xc9(201)
Neighbor Interface is fc1/7 (0x00010006)


bdc-mds9506-2# show fspf database vsan 1
FSPF Link State Database for VSAN 1 Domain 0x1f(31)
LSR Type                = 1
Advertising domain ID   = 0x1f(31)
LSR Age                 = 1695
LSR Incarnation number  = 0x8000023e
LSR Checksum            = 0xa174
Number of links         = 2
 NbrDomainId      IfIndex   NbrIfIndex    Link Type         Cost
-----------------------------------------------------------------------------
   0xc9(201) 0x00010007     0x00010006               1          250
   0xc9(201) 0x0001000b     0x00010004               1          250


To see actual routing table for unicast:

mds9506# show fcroute unicast

D:direct  R:remote  P:permanent  V:volatile  A:active  N:non-active
                                                          # Next
Protocol  VSAN       FC ID/Mask        RCtl/Mask    Flags  Hops  Cost
--------  ----    -------- --------    ---- ----    -----  ----  ----
local     1       0x1f0000 0xffffff    0x00 0x00    D P A  1     1
fspf      1       0xc90000 0xff0000    0x00 0x00    D P A  2     250
fspf      100     0x8c0000 0xff0000    0x00 0x00    D P A  2     250

? show system internal rib unicast

Port Channels

Port channeling allows to aggregate multiple links into one logical interface to provide higher aggregated bandwidth, load balancing and link redundancy.

Cisco MDS switch supports two types of Port Channels: Fibre Channel and Ethernet.

Ethernet Port Channels group together Gigabit Ethernet interfaces on the IPS modules. It is similar to the port channels on LAN words, except:

1. don't support PAgP and LACP protocols. The port channel must be configured statically
2. can only aggregate two physical interfaces adjacent to each other on the IPS module (1-2, -4, 5-6 or 7-8)
3. are supported on IPS-4 and IPS-8 modules. They are not supported on MPS14/2 and 9216i dual GigE ports.
4. If you are running a single FCIP tunnel on an Ethernet Port Channel, only one of the two links will be used under normal conditions, because of the load balancing on the Cat Ethernet switches. There's a 50/50 chance that IPS module will use one link for outbound traffic, and Cat Ethernet switch will use the other link for inbound traffic.

Fibre Channel port Channels group together FC or FCIP tunnel interfaces:

1. Up to 16 interfaces across different slots can be grouped together
2. MDS supports up to 128 port channels with a mix of Gen-1 and Gen-2 modules
3. MDS supports up to 256 port channels when only Gen-2 modules are installed in a switch
4. Each group of physical links appears to FSPF (Fibre Channel Routing protocol) as one logical link. 
5. Only E and TE ports can be port channeled. F ports cannot be port channeled (HBAs and Storage don't support it)
6. Port Channels between Cisco and non-Cisco switches are not supported
7. Load balancing is performed per VSAN based on the VSAN parameter: vsan 1 loadbalancing [src-dst-id | src-dst-ox-id]. By default, load balancing is Exchange based (src-dst-ox-id) and can be changed to Flow based (src-dst-id). The same VSAN load-balancing parameter applies to Port Channels and equal-cost FSPF paths.

Port Channel protocol uses EPP (exchange peer parameters) SW_ILS frames. SW_ILS fames are special Fibre Channel frames exchanged between storage switches on ISL links. There are mane kinds of SW_ILS frames and EPP is one of them.

Port Channel protocol consists of two sub-protocols:

1. Bring-up Protocol. When Port Channel comes up, two peering switches exchange Port Channel configuration. If the settings don't match, the Port Channel  individual interfaces go into an isolated mode (port channel up, but doesn't pass any data). As soon as settings match on both ends, the Port Channel will become operational. The following settings should match at both ends of the Port Channel:
1. Port Type (FC or FCIP)
2. Configured Port mode (AUTO or E)
3. Trunk mode (ON, OFF or AUTO)
4. Port Speed 
1. Generation-1: Auto, 1000 or 2000
2. Generation-2: Auto, Auto MAX 2000, 1000, 2000 or 4000
3. Generation-3: Auto, Auto MAX 2000, AUTO MAX 4000, 1000, 2000, 4000 or 8000
5. Port VSAN
6. Trunk allowed VSAN list. Make sure that VSANs are first created on both switches before trying to use switchport trunk allowed vsan command. If a VSAN does not exist, the Port Channel will be in the isolated mode.
7. Port Security
8. FibreChannel receive buffer size 
9. This list can be viewed by running show port-channel compatibility-parameters command
2. Autocreating protocol. This protocol makes sure that ports are eligible to the part of the same port channel and that they have an eligible partner switch. It then automatically aggregates these ports into a Port Channel. 
1. This feature must be enabled manually for each individual interface
2. SAN-OS will start with number 128 for the auto-created port channels, and go down 127,126, etc. (if all modules are Gen-2, it will start with 256)
3. an autocreated Port Channel number may be different across reboots
4. If some interface doesn't join the port channel automatically, that is because its settings don't match the other ports in the autocreated port channel. That interface will come up as a separate link.
5. Any configuration changes made to the autocreated port channel are not saved under the autocreated port channel interface. Instead they are only saved under the individual interfaces that are part of the port channel. Some configuration settings cannot be made for the autcreated port channels.
6. You can convert autcreated port channel into a normal port channel
7. both sides of an autocreated port channel must be autocreated.

MDS Fibre Channel port Channels have two modes:

1. Port Channel ON mode. 
1. (conf-int)#no channel mode active
2. port channel protocol is OFF
3. you have to explicitly no shut interfaces after making them part of a port channel
4. mode is default
5. suspends interfaces if their operational values are incompatible with Port Channel
2. Port Channel ACTIVE mode
1. (conf-int)#channel mode active
2. Port Channel protocol is ON
3. port channel automatically bundles additional available interfaces
4. dynamic link recovery when adding or deleting links
5. isolates interfaces if their operational values are incompatible with Port Channel

It is not required to have the same mode on both ends of the port channel. One side can be ON (default) and the other side can be ACTIVE.

There are tree ways to create port channels:

1. configure each individual port or range of ports with channel-group X or channel-group X force command
1. Port channel X interface will be automatically created
2. port channel mode will be set to ON 
3. individual port members will be shut down and you'll have to no shut them manually
2. configure port channel interface first with interface port-channel x command. 
1. Port channel mode will be set to ON, but you can change it to ACTIVE mode by using channel mode active.
2. configure each individual port or range of ports with channel-group x or channel-group x force command.
3. Configure each individual port or range of ports with command channel-group auto and no shut them.
1. SAN-OS will use number 128 (256 for all Gen-2 modules) for the first autocreated port channel
2. you won't see port channel 128 in the configuration, but you can still conf t, int po128 and configure some of its settings. These settings will be saved to the individual port channel member interface configuration.

The force keyword on the channel-group command plays a role only when adding ports to the port channel after the port channel was already created.

We can't modifu FSPF parameters on the auto creatd port-channels. We must first convert it to a user-configured port channel:
#port-channel 128 persistent

VSANs

VSANs isolate servers and storage devices that are physically connected to the same fabric, therefore allowing creating of logical SAN islands on a shared hardware. Each VSAN effectively functions as a separate SAN. VSAN is a Cisco proprietary feature.

Each VSAN has its own fabric services completely isolated from fabric services in other VSANs. For instance: FCID Domain Manager, FCNS Name Server, FSPF, Zoning, Port Security database. Fabric Events are isolvated per-VSAN basis.

VSANs have the same purpose as VLANs.

Each VSAN has its own:

1. Domain IDs and Principal Switch priority
2. Node FCIDs
3. Zones and Zonesets
4. FSPF routing topology
5. Load-balancing method across FSPF Equal Cost paths and across FC-Port Channel individual links
6. Time-out values
7. in-order-guarantee setting
8. Interop mode. This feature is used when other non-Cisco storage switches are connected to a port in a VSAN.

A maximum of 256  VSANs can be configured, but allowed VSAN IDs are from 1 to4093. VSAN 1 is default and cannot be removed. VSAN 4094 is always isolated and is used for those ports for which VSAN was just deleted. These ports should be manually reassigned to an active VSAN.

I have to configure VSANs manually on each MDS switch. CFS does not distribute VSANs between switches. I am not required to configure each VSAN on every MDS switch. If an MDS switch has no FC interfaces in a given VSAN and will not carry any transit traffic for that VSAN then I don't need to configure it on the switch.

Port VSAN membership on the switch is assigned on a port-by-port basis either manually (by configuring VSAN database on  a switch) or dynamically (Dynamic Port VSAN Membership=DPVM). SAN-OS checks the DPVM database during a device FLOGI and assigns it to a specific VSAN based on the device port's pWWN or nWWN. DPVM can be distributed between MDS switches with CFS. DPVM is not supported on TL and FL ports. It is only supported on F ports. Dynamy Port VSAN Membership (DPVM) is also referred to as WWN-Based VSANs or Dynamic VSANs.

When activating PDVM database, I should always use the force keyword.Otherwise, if nWWN or pWWWN entry in the database overlaps with an nWWN or pWWN of an aready active N_Port on any of the switches, the dpvm commit will fail.

Show vsan membership does not reflect DPVM configuration, this command displays ony static VSAN assignments. Instead, use show interface brief - it displays the active VSAN ID for each port whether it was assigned statically or dynamically.

BB_Credits

Fibre Channel devices employ a credit-based flow control mechanism to guarantee delivery between storage devices and to ensure that the rate at which the data i sent by the source does not exceed that at which it can be received at the destination.

To establish a link-be-link flow control, ports of the two connected storage devices first exchange and agree on the number of frames each can receive (server HBA and switch port, for example). This is called the buffer-to-buffer credit (BB_Credits). Whenever the source device sends out a frame, it increments the credit counter by one. The receiving device will send back an acknowledment message R_RDY primitive upon the receipt of each transmitted frame. Once the source device receives the R_RDY, it lowers the BB_Credit counter by one. If the BB_Credit counter reaches the previously agreed upon credit threshold, the source device simple stops sending frames until the BB_Credit counter is lowered below its threshold by receipt of another R_RDY from the destination.

The same behavior happens in another direction. BB_Creding max value can be different on two ends of an FC link.

Default receive BB_Credints depend on the module. For 16 port module max is 255 (Fx_Port default is 16, and E_Port default is 255). For the 32-port linecard, it is fixed at Fx_Port 12 and E_Port 12.

Transmit BB_Credits are automatically set to the advertised receive BB_Credits of the peer port. This is done during FLOGI process.

Receive Performance buffers (no transmit) are on by default. They range from 1 to 145. They are used in addition to the BB_Credits. SAN-OS uses build-in algorithm to calculate how many performance credits are needed. On E_Port it is typically at 145 and on F_Port it is typically 12.


mds9124(config)# show int fc1/3 bbcredit
fc1/3 is up
    Transmit B2B Credit is 3
    Receive B2B Credit is 16
      16 receive B2B credit remaining
      3 transmit B2B credit remaining
      3 low priority transmit B2B credit remaining

BB_Credits can be different on two links in a Port Channel. BB_Credits affect only individual inks.

CLI variables and command aliases

You can create CLI variables that could be referenced in the CLI commands:


mds9124(config)# cli var name start running-config

mds9124(config)# cli var name run startup-config

mds9124(config)# cli alias name wr copy $(start) $(run)

mds9124(config)# wr

[########################################] 100%

Copy complete, now saving to disk (please wait)...

Device alias database

When a fibre channel disk and fiber channel HBA ports join the fabric, they send a FLOGI frame (Fibric Login). The fabric (MDS switch) then assigns an FCID to each fibre channel disk and each fibre channel HBA port. These FLOGI mappings of pWWWN/nWWN to FCID are stored in the FLOGI database.

In order to configure device alias database, I need to know pWWWNs of the HBAs and JBOD drives. show flogi database could tell me pWWWNs of them:


mds9124# show flogi database
--------------------------------------------------------------------------------
INTERFACE        VSAN    FCID           PORT NAME               NODE NAME
--------------------------------------------------------------------------------
fc1/3            1     0xc90000  21:00:00:e0:8b:0a:fc:6d 20:00:00:e0:8b:0a:fc:6d

Total number of flogi = 1.

3rd, 4th and 5th bytes represent OUI (if you look up the OUI value 00:e0:8b on http://www.coffer.com, you could confirm that the HBA is made by Qlogic).

Think of each drive as an individual NL_Port on the Fibre Channel Arbitrated Loop. MDS port connected to HBA must run on TL mode.

The device-aliases have global significance, therefore, we could configure device-alias on one MDS, and later on (when we enable links between MDSes) the device-alias database will be distributed to other MDSes automatically using CFS CW_ILS frame.

We could have two different mode for hte device-alias: basic and enhanced. Enhanced mode is recommended if you are configuring feature SDV(SAN Device Virtualization). In basic mode, device-aliases are referred to by pWWN in the zoning configuration. In enhanced mode, device-aliases are referred to by name. Therefore, if pWWN changes, you don't need  to update zoning configuration.


mds9124# show device-alias status
Fabric Distribution: Enabled
Database:- Device Aliases 0  Mode: Basic
           Checksum: 0xf6bd6b3389b87233d462029172c8612

mds9124# conf t
Enter configuration commands, one per line.  End with CNTL/Z.
mds9124(config)# device-alias mode enhanced
mds9124(config)# show device-alias status
Fabric Distribution: Enabled
Database:- Device Aliases 0  Mode: Basic
           Checksum: 0xf6bd6b3389b87233d462029172c8612
Locked By:- User "CLI/SNMPv3:admin" SWWN 20:00:54:7f:ee:e8:25:80
Pending Database:- Device Aliases 0  Mode: Enhanced

The device-alias configuration is locked. In order to do it active, we should commit the change to CFS:

mds9124(config)# device-alias commit

mds9124(config)# show device-alias status

Fabric Distribution: Enabled

Database:- Device Aliases 0  Mode: Enhanced

           Checksum: 0xb713791d47fd2d205e1a5fa01ea9c447

To change mode to basic:

mds9124(config)# no device-alias mode enhanced

mds9124(config)# device-alias commit

To configure device alias for the Initiator:

mds9124(config)# show device-alias database

There are no entries in the database

mds9124(config)# device-alias database

mds9124(config-device-alias-db)# show flogi database

--------------------------------------------------------------------------------

INTERFACE        VSAN    FCID           PORT NAME               NODE NAME

--------------------------------------------------------------------------------

fc1/3            1     0xc90000  21:00:00:e0:8b:0a:fc:6d 20:00:00:e0:8b:0a:fc:6d

Total number of flogi = 1.

mds9124(config-device-alias-db)# device-alias name Initiator pwwn 21:00:00:e0:8b:0a:fc:6d

mds9124(config-device-alias-db)# exit

mds9124(config)# show device-alias database

There are no entries in the database

mds9124(config)# device-alias commit

mds9124(config)# show device-alias database

device-alias name Initiator pwwn 21:00:00:e0:8b:0a:fc:6d

Total number of entries = 1

mds9124(config)#

Because my link between mds9124 and mds9506 is up, without any further configuration, we got:

mds9506# show device-alias database

device-alias name Initiator pwwn 21:00:00:e0:8b:0a:fc:6d

Total number of entries = 1

Fcaliases are not supported by CFS distribution, therefore you have to configure them on every MDS manually. Fcaliases have logical scope (VSAN-based), support multiple members, you can use identifiers other than pWWN:

mds9124(config)# fcalias name TestFcalias vsan 1

mds9124(config-fcalias)# member ?

  device-alias       Add device-alias member to fcalias

  domain-id          Add member based on domain-id,port-number

  fcid               Add FCID member to fcalias

  fwwn               Add Fabric Port WWN to fcalias

  interface          Add member based on interface

  ip-address         Add IP address member to fcalias

  pwwn               Add Port WWN to fcalias

  symbolic-nodename  Add Symbolic Node Name to fcalias

mds9124(config-fcalias)# member pwwn  50:0a:09:82:99:9b:48:e0 ?

  <CR>

  both       Enter device-type as both

  initiator  Enter device-type as initiator

  lun        Add PortWWN-LUN member to fcalias

  target     Enter device-type as target

mds9124(config-fcalias)# member pwwn  50:0a:09:82:99:9b:48:e0

mds9124(config-fcalias)# show fcalias vsan 1

fcalias name TestFcalias vsan 1

  pwwn 50:0a:09:82:99:9b:48:e0 [Initiator]

Since fcalias does not support CFS, we do not have this information on the second MDS:

mds9506# show fcalias vsan 1

Alias not present

We can import fcalias into device-alias database:

mds9124(config)# device-alias import fcalias vsan 1

mds9124(config)# device-alias commit

Now we do see this information on seconds MDS:

mds9506# show device-alias database

device-alias name Initiator pwwn 21:00:00:e0:8b:0a:fc:6d

device-alias name TestFcalias pwwn 50:0a:09:82:99:9b:48:e0

Total number of entries = 2

CFS = Cisco Fabric Services

CFS is a SAN-OS feature indtroduced in 2.0(1b) release. It allows you to perform almost any configuration tast on one MDS switch and automatically copy it to other MDS switches in the fabric using fibre-channel connections. Not al appications support CFS.

You could see which applications are supported and enabled by CSF by using the command: show cfs application:

mds9506# show cfs application

----------------------------------------------
 Application    Enabled   Scope
----------------------------------------------
 ntp            No        Physical-fc-ip
 fscm           Yes       Physical-fc
 role           No        Physical-fc-ip
 rscn           No        Logical
 radius         No        Physical-fc-ip
 fctimer        No        Physical-fc
 syslogd        No        Physical-fc-ip
 callhome       No        Physical-fc-ip
 fcdomain       No        Logical
 fc-redirect    Yes       Physical-fc
 device-alias   Yes       Physical-fc

Total number of entries = 11

Physical connection

Did my first lab setup. Put two connections between two MDS switches + put connection to the initiator. Had several problems and learned the following:
1. some MDS platforms have license per group of ports.
2. Fiber cable has a light on one of the channel. This channel we should plugged in into dark side of SFP.
3. We need to check which SFP type is good for the platform. The matrix could be found at http://www.cisco.com/en/US/prod/collateral/ps4159/ps6409/ps4358/product_data_sheet09186a00801bc698.html


So I have the following topology:

Initiator--- [fc1/3]mds-9124 [fc 1/5][fc1/7]--- [fc 1/8][fc1/12] mds-9506 [fc1/4]--Storage (netApp)

TODO plan for the week 20-27th of May

Had meeting with mentor. Some ideas, what I should test in the lab:

1. Zoning, zoning with aliases, if more than 1 user is doing zoning at the same time, then ...
2. VSANs
3. fc/device aliases
4. Zone default policy, enchencement zoning
5. fcns database details
6. how to debug flogi/plogi - show system internal flogi event???
7. How to find the path from source to the target: show flogi database interface, show fcns database <name server>, show topoloy, FSPF
8. FSPF - like OSPF in SUN word
9. show logging log
10. show accounting log
11. For 9500 - check how system switchover is working between Sups
12. Change LUN slot into NPIV mode. What is the difference between NPV and NPIV?
13. FCIP - for this I need a special module. This is optional for the next week