| A | B |
|---|
| vPC definition | (1)eliminate STP blocked ports and uses all available BW(2)port channeling concept allows link aggregation between 2 switches(3)seamless VM mobility. Server HA clusters(4)scale L2 BW(5)independent CP(6)fast convergence(7)simplify net design(8)all models(9)etherchannel LB hash |
| vPC domain | A pair of vPC switches - must be unique for all vPC pairs defined within contiguous L2 domain - NOTE: domain ID is used to create the system MAC on the switch |
| vPC peer | A vPC switch, one of a pair |
| vPC member port | one of a set of ports (port channels) that form a vPC |
| vPC | the combined port channel between vPC peers and downstream device |
| vPC peer-link | link used to synchronize state between vPC peer devices must be 10G-must be a trunk - carries CFS - carries flooded traffic from vPC peer - carries BPDUs, HSRP Hellos, IGMP upgrades etc. - PTP - 7K Fibre Card - any 10G on 5K - recommended to have 2 10G - should be dedicated |
| vPC peer-keepalive-link | link between vPC peer devices - keepalive not as stringent as the keepalive is not used unless there is a failure - if is a failure on peer-link but keepalive up Secondary will suspend ALL ports |
| vPC VLAN | any of the VLANs carried over the peer-link & used to communicate via vPC with a peer device |
| non-vPC vlan | any of the STP vlans not carried over the peer link |
| CFS | Cisco Fabric services protocol - used to synchronize and config validation between peer devices |
| Data Plane Loop Avoidance | orphans traffic can traverse peer-link all other traffic is marked and not allowed to egress on a vPC |
| Config Steps | (1)define domains(2)establish peer keepalive connectivity-(L3 connnectivity)(3)create peer link(4)reuse port-channels and create vPCs(optional) |
| vpc verification commands | (1)sh vpc role |
| Role Priority | primary process BPDUs - secondary suspends ports when there are configuration issue |
| vPC consistency checks 2 types | (1)type 1 graceful - only ports on secondary switch suspended (2) type 2 - Error messages to indicate potential of undesired behavior - system consistency must be kept in sync |
| VDCs | each VDC should have its own vPC domain - separate vPC peer-link andf peer keepalive infrastructure for each VDC deployed - using 2 VDCs on same 7K to for a vPC domain is not supported |
| vPC configuration | (1)feature vpc (2)feature lacp(3)vpc domain 10(4)peer-keepalive destination 1.1.1.2 source 1.1.1.1(5)system-priority 3000(6)role priority 90(7)peer-gateway(best practice to enable)(8)int po1(9)switchport mode trunk(10)vpc peer-link(11)spanning-tree port type network |
| downstream vpc member | (1)int po201(2)switchport mode trunk(3)switchport trunk allowed vlan 100-105(4)vpc 201 |
| enhanced vpc | vpc to fex then from fex to server - supported from 5Ks not 7K |
| L3 and VPC | hellos cannot go over VPC port channel |
| FCoE and VPC | basically one side is for Ethernet and one side is for SAN/Ethernet |
| vPC auto-recovery | (1)vpc domain 10 (2) auto-recovery |
| steps for vPC failure | (1)secondary if peer-link goes down it shuts down all its ports (2)if primary goes down after 3 keepalives it becomes the primary and begins bringing up vPCs (3)if auto-recovery is configured it allows new links to be brought up doesn't allow flapping |
| vPC orphan suspend | prior to release 5.2 orphan port was not shut down on loss of vPC peer link - you can configure the secondary peer to shutdown thus triggering NIC teaming recovery for all teaming configs - config applied to physical port (1)int e100/1/1(2)vpc orphan-ports suspend |
| arp synchronization | needs to be enabled on both switches - when peer-link comes up it does a bulk sync over CFSoE to peer switch - improve convergence for L3 flows (config-vpc-domain)# ip arp synchronize (not enabled by default) |
| vPC delay restore | delays L3 restore giving L2 interfaces time to come up - allows for more graceful L3 restore - is enabled by default 30 second interval |
| vPC peer-switch | allows both switches to be seen as a root - both switches send same BPDU out |
| 7K | cannot do eVPC - cannot VPC between 2 chassis |
| Traits of FabricPath | (1)up to 256 links (2)ECMP(3)shortest path(4)single ingress lookup(5)enhanced L2 only works on F1 and F2 mods and 5500 NOTE: need enhanced L2 license |
| FabricPath topology | FabricPath - Spine & FabricPath Leaf - uses SPF (shortest path tree) between switches - each switch is given a FabricPath ID - runs CLNS - does not run on IP |
| FabricPath How it works | (1) ARP Req (2) Encapsulated in IP (3) Encapsulated in Ethernet (4) Encapsulated in FabricPath - floods this information everywhere much like spanning-tree - then ARP reply comes back - within FabricPath header is Src Switch ID and Dst Switch ID though |
| FabricPath routing | routes based on Switch ID - Spine does not learn MAC address table only Switch ID table (aka Outer Destination Addresss 48 bits - Outer Source Address 48 bits) - otherwise known as TRILL - can have L2 or L3 FabricPath in the core |
| parts to Classic FabricPath Pod | (1)simple configuration - no peer link - no switch pairs - no port channels (2)design flexibility - easily extensible (3) No STP - no traditional bridging - no topology changes - no loops |
| parts configuring FabricPath | (1) admin VDC install feature-set fabricpath(2)w/i VDC turn on feature-feature-set fabricpath(3)fabricpath switch-id 10(4)spanning-tree vlan 6-20 priority 8192(5)add FP interfaces(config-if)# switchport mode fabricpath(6)add FP vlans(config-vlan)# mode fabricpath |
| parts to add vPC+ configuration | (7)config-vpc domain 1(8)(config-vpc-domain)#fabricpath switch-id 1000 |
| FabricPath and IS-IS | IS-IS maintain FP routing table using switch ID |
| FabricPath and Conversational MAC | once is enabled conversational MAC learning is enabled on all core devices - not enabled for CE though - all VLANs are in CE by default |
| FabricPath commands | (1)show cdp neighbors(2)show fabricpath isis adjacency(3)show fabricpath switch-id(4)fabricpath isis topology 0 route(5)show fabricpath ftag(6)show mac address-table |
| Ftag | forwarding tags - determines how a packet is forwarded over fabricpath topology |
| fabricpath topology | default topology 0 can configure from 1 - 63 |
| Hardware profile verification | show hardware profile status |
| L2 switching troubleshooting | (1)show interface e1/1 ? - brief - capabalities - counters - transcenciever - trunk - status - switchport - mac-address (2)show mod (shows mac-addresses - aka bia) |
| very detailed interface status | show hardware internal info port-info valid-only (up & running if use valid only) - NOTE: on some switches you must auto set if is below 10G |
| spanning tree information verify | Show spanning-tree vlan 123 |
| show system internal ethpm info e1/1 | ethpm - takes care of port manager events - gives all kinds of information |
| sh sys intern ethpm event-his int e1/1 | gives a list of events on the interface - if there are a lot you can use the last or Include or grep |
| unicast/multicast MAC table | (1)show mac address-table [dynamic|multicast] - software cached entires (2)show platform fwm info hw-stm - hardware entries (3)show int vlan 200(4)show hard intern bcm-usd info tables l2 l2-mod-fifo all slot-num 0 |
| control traffic to CPU | (1)show hardware internal bcm-usd info tables l2 l2-user-entry-only all slot 0 [valid-only](2)show hardware internal bcm-usd info tables l2 l2-user-entry-data-only all slot 0 |
| mac aging timer | mac address-table aging-time 1000 vlan 100 - certain platforms is not a per-vlan aging time |
| FCoE defined | (1)map FC frames over Ethernet (2)enables FC on a lossless ethernet Network |
| FCoE Benefits | (1)Fewer cables - both block I/O & Ethernet traffic on same cable (2)fewer adapters needed (3)overall less power(4)interoperates with existing SANs-management of SANs remains constant(5)no gateway |
| FCoE T11 | FC-BB-5 |
| FCoE 802.1 DCB encompass all 3 | (1)PFC - lossless ethernet 802.1Qbb (2)ETS - priority grouping 802.1Qaz (3)DCBx 802.1Qaz |
| FC Buffer-to-Buffer Credits | R_RDY receives from target and then the initiator Transmits frame |
| PFC | VLAN Tag enables 8 priorities for Ethernet traffic (2)PFC enables Flow Control on a per-priority basis (3)Therefore we have the ability to have lossless and lossy priorities at same time on same wire-allows FCoE to be lossless independent of other priorities |
| FCoE configuration | (1)feature fcoe - 2 classes are made by defaulkt (2)class-fcoe is configured to be no-drop with an MTU of 2158 (3)best practice is to use default CoS value of 3 for FCoE/no drop(4)class type network-qos class-fcoe (5)pause no-drop (6)mtu 2158 |
| verfication of cos config | show class-map |
| copp policy | show policy-map interface control-plane |
| priority flow control verification | show interface priority-flow-control (have an 8 there means have PFC) |
| ETS | enhanced transmission services - (1)allows you to create priority groups (2)can guarantee BW (3)can assign BW percentages to groups (4) not all priorities need to be used or in groups (default and best practice is 50% tp FCoE and Ethernet |
| DCBx - data center bridging exchange | devices need to discover edge of enhanced Ethernet cloud-each edge switch needs to learn it's connected to legacy switch-servers need to learn whether or not are connected to enhanced ethernet device w/i enhanced ethernet cloud, devices need to discover capabilities of peer |
| DCBx - data center bridging exchange | uses LLDP to get information |
| FCoE | Fibre Channel frame wrapped in ethernet header(2)when hits Fibre Channel rip off Ethernet and FCoE header and are good to go |
| FCF | Fibre Channel Forwarder (Nexus 5K, 7K, 9K) - does the FLOGI for FCoE - is the FC switching element within FCoE swith - consumes a Domain ID - FCoE encapsulation and decapsulation occurs here - can handle both FC and FCoE - FLOGIs |
| Enode | FC node that is able to transmit FCoE frames using one or more Enode MACs - CAN |
| Enode MAC | lossless ethernet MAC coupled with an FCoE controller in an Enode |
| FC-MAP | Fabric Provider MAC address - required value for the upper 24 bits of a MAC addressed assigned VN_Port |
| FCF-MAC | Lossless ethernet MAC coupled with FCF - is FC-MAP plus FC-ID aka domain id - FC-MAP=upper 24 bits of Enode's MAC | FC-ID=lower 24 bits of Enode's MAC |
| FPMA | unique MAC address assigned by FCF to a single Enode - a fabric provided MAC Address associated with a single VN_Port at that Enode |
| FCoE Pass Through | DCB device capable of passing FCoE frame to an FCF via FIP snooping |
| Multi-Hop FCoE | extension of FCoE beyond a single hop |
| Adapter FEX | allows you to go to server level and map multiple virtual interfaces over a singe ethernet interface using "channels" |
| Adapter FEX Topologies | (1) single homed (straight to 5500) (2) single-homed 2k to 5500(3)dual-homed - 2 5500s to 1 FEX(4)Active-Standby - 2 nexus 5k(5)Active-Standby with FEX |
| Adapter FEX config (auto & manual) | (1)(config)#install feature-set virtualization(2)(config)#feature-set virtualization(3)(config)#veth auto-create(4)port-profile type vethernet user_data(5)(config-if)#switchport mode vntagMANUAL-(6)(config)#vethernet 21(7)(config-if)#bind int e101/1/15 channel 1 |
| Adapter FEX config for 2K 11 ( auto/ man) | (1)install feature-set virtualization(2)feature-set virtualization(3)fex 101(4)fcoe(3)veth auto-create(5)int e101/1/1(6)switchport mode vntagMANUAL-(1)vethernet 21(2)bind int e101/1/1 channel 1(3)switchport mode trunk (4)int vfc 4(5)bind int vethernet21 |
| Adapter FEX peer 2k config | (1) different fex # (2) different channel (3) different vethernet (4)different vfc |
| Adapter Fex FCOE | (1)single physical link split into multiple virtual channels(2)channels(3)FCoE switches 5500 and 2232(4)FCoE server UCS P81E for C-Series, support VNTag like BCM57712 |
| Adapter FCoE channels | (1) identified by unique channel number(2)channel scope limited to physical link(3)Connects a server vNIC with a switch vEthernet interface(4)uses tagging with VNTag identifiers |
| Fibre channel frame format | (1)Word (2)Frame (3) Sequence (4) Exchange |
| Fibre Channel Config | (1)interface fc2/1-5 (2)switchport mode E (3)switchport mode auto(4)switchport fcrxbufsize 2000(5)switchport fcrxbbcredit 5 mode e(6)no system default switchport shutdown san - turns default fc setting to up instead of down (default) |
| RCSN | registered state change notification anything changes this is sent out - disks leaving, entering etc. - SW-RCSN - sent switch to switch |
| N_Port communication steps | N_Port logs into its attached F_Port - FLOGI or fabric login |
| FLOGI | Fabric Login - pWWN (hardware-HBA) and get a FCID (logical-assigned) - this is how stuff is "routed" in the Fibre Channel World - show flogi database - shows all the mappings |
| PLOGI | N port logging into to its target N port |
| PLRI | N port must exchange ULP with target to ensure target and initiator can communicate |
| pWWN | port world wide name - identify port in device - 64 or 128 bits |
| nWWN | node world wide name - identify device - 64 or 128 bits |
| Fibre Channel Address Format | (1)Domain - defines a switch - 8 bit field only 239 allowed (2)Area - groups of ports within domain (3)Port-ID - devices on port |
| FCID | first byte is a Fibre Channel Domain (akin to a SM in the ip world) |
| VSAN | is like a vlan - is carried on a TE port or trunk expansion port |
| F Port | Fabric Port Access Port - actual port on switch -always connects to N - note expects only one host |
| N Port | Node Port - CNA - SAN - UCS etc. - end user port - anything with an HBA - target or initiator - connects only to F |
| E Port | Expansion Port - ISL - port connecting 2 switches together - aka dot1q - must hardset it (1) (config)# fc2/13 (2) (config-if)# switchport mode e - some autonegotiate this - only goes E to E or switches |
| NP Port | an N Port in NPV mode connected to a switch via and F_Port (has multiple logins is a "hidden switch") |
| TE Port | trunking extension port - create EISL between switches - multiple VSANs |
| TF Port | trunking f port expands functionality of F ports to support VSAN trunking - connects to TNP |
| TNP Port | connects to a Trunked F port or TF port |
| SD Port | Spanned Destination Port |
| VSAN trunking configuration | (1)int fc2/2(2)switchport mode e(3)switchport trunk mode on(4)switchport speed 4000(5)switchport trunk allowed vsan 1-10 |
| VSAN numbering | (1)VSAN 1 - default (2) VSAN 2 - 4093 - user configurable (3) 4094 - isolated vsan |
| VSAN traits | (1)VSAN separate routing and namespance(2)limit uni,multi,broadcast traffic(3)members - physical port or pWWN(4)endpoint - HBA(5)member enforcement-at each E port,source and dest port(6)scope-large(7)config changes-only when ports needed(8)use-per app or dept |
| Zone traits | (1)zone in same vsan same routing(2)limit unicast traffic(3)members - pWWN(4)endpoint - HBA can be in multiple(5)member enforcement-source and dest port(6)scope-initiator and target not outside zone(7)config changes-frequent(8)use-single initiator |
| NPV | node port virtualization-extension to NPIV- allows blade switch or ToR fabric device to behave as an NPIV based HBA to the core Fibre Channel Switch-aggregates host ports(N ports) into one or more uplinks (psuedo-interswitch links)to core switches - allows to save domain IDs |
| NPV explained parts | (1)FC services-most switched off(2)switching operation-acts as proxy - subordinate to FC switch(3)does not use Domain ID - no domain ID limitation(4)scalability and manageability-eliminates need for server adminstrators to manage SAN-3rd party integration VSAN scalable(5)no QoS |
| NPV Mode | (1)NPV edge switch aggregates locally connected host ports (N Ports) into one or more uplinks to the core switches (2)allows blade and TOR switches to behave as NPIV-based HBA to core Fibre Channel Switch |
| NPV config | (1)feature NPV(2)vsan database(3)vsan 3 int fc2/1,fc7-8(4)vsan 3 interface fc2/2, fc9(5)int fc2/1(6)switchport mode F(6)int fc2/8(7)switchport mode NP(goes to NPIV switch)(8)npv traffic-map server-int fc 2/1 external-interface fc2/7 (manually pinning) |
| verifying NPV | (1)show npv status (2)show npv traffic-map |
| FCoE for NPV license | (1)Storage Protocols Services Package is needed - if do feature fcoe and feature npv- do wr erase and reboot switch (2) if do feature fcoe - npv does not do write-erase requires fcoe_npv_pkg |
| fcoe NPV config | (1)feature fcoe-npv(2)feature lacp(3)qos (optional)(4)vsan database(5)vsan 5(6)vlan 50(7)fcoe vsan 50(8)int vfc 1(9)bind int e2/1(10)switchport mode F(11)int vfc 130(12)bind int po13(13)switchport mode NP |
| verification fcoe npv | show int vfc 1 - show int vfc 130 - show vlan fcoe - show npv status - |
| NPIV | node port ID virtualization - (1)provides a means to assign multiple FCIDs to a single N port(2)multiple applications can use same HBA(3)use of different pWWNs(4)usage applies to virtual servers-VMWare,HyperV,Xenserver |
| NPV and NPIV support Core/Edge Models | Edge -NPV&NPIV (1)MDS 9124,9134,9148(2)Fibre Channel Blade switches IBM and HP(3)Nexus 5K and 5500(4)UCS 6100 and 6200 -Core-NPIV-(1)MDS 9500 Series MultiLayer Directions(2)MDS 9216 multilayer and 9222i(3)MDS 9124,9134,9148(4)3rd party switches |
| NPIV config | (1)feature NPIV(2)int fc1/1(3)switchport mode F(4)switchport mode F(5)switchport mode trunk - NPIV F port prepped for multiple requests |
| verification npv & npiv | show [npv|npiv] database |
| CNA | converged network adapter appears to OS as separate HBA(FC Drivers) and NIC (Ethernet Drivers)- CAN supports DCB, PFC and FIP |
| FIP frames | (1)FIP vlan discovery(2)FIP discovery(3)FCoE virtual link establishment(4)FCoE Virtual Link maintenance(5)FIP done per-VLAN(6)recommended to use FIP VLAN discovery protocol on default VLAN(7)all FIP protocols performed in each vlan w/FC-BB_E Services |
| FIP multicast addresses | (1) 01-10-18-01-00-00 = group address for all FCoE devices (2) 01-10-18-01-00-01=group address for all ENodes, used by multicast discovery advertisements (3) 01-10-18-01-00-02=group address for all FCFs used by VLAN discovery request, multicast discovery solicitation |
| FCoE Protocol Addressing Steps | (1)LLDP DCBx from Host(note TLV gets all specifics)(2)LLDP DCBx from Switch(3)FIP VLAN request from MAC(H1)to all FCFS-MAC(4)FIP VLAN response from FCF-MAC to MAC(5)FIP discovery soliciation(6)FIP discovery advertisements |
| FCoE verification commands | show fc2 internal fcoe - show int mgmt (observe MAC) |
| NPV | do not want to consume FC Domain ID, zone etc for each device so the devices login to an NPV or NPIV - NPV converts FLOGI to FDISC - Fibre Discovery |
| NPV verification commands | (1)show fcns database detail |
| FCoE NPV method 1 | Method1 (1)enable FCoE feature then NPV(2)Req. full storage services license(3)wr erase reload done(4)method allows FCoE & FC upstream & host NPV connections erase(4)no storage license |
| FCoE NPV method 1 configuration | (1)feature fcoe (2) feature npv |
| FCoE NPV method 2 | Method2(1)enable fcoe-npv(2)preferred if have pure ethernet environment with FCoE(3)no wr erase - if you do not need FC |
| FCoE NPV method 2 configuration | (1)feature fcoe-npv - once again you can use if there is no FC involved and is pure ethernet |
| N7K Feature Enabling | (1)install feature-set fcoe(2)feature lacp(3)feature lldp(4)license grace-period(5)license fcoe module 4 |
| MDS | (1)install feature-set fcoe(2)feature-set fcoe(3)feature lldp(4)feature vlan-vsan-mapping(5)feature lacp(6)feature npiv - 1st 4 are done automatically |
| 7K QoS policy for FCOE | (1)system qos (2)(config-sys-qos)#service-policy type network-qos default-nq-7e-policy(3)also need to create a storage VDC |
| 5k QoS policy | (1)system qos (2)(config-sys-qos)#service-policy type qos input fcoe default-in-policy(3)service-policy type queuing input fcoe default-in-policy(4)service-policy type queuing output fcoe default-out-policy(5)service-policy type network-qos fcoe-default-nq-policy |
| QoS and 5K | should automatically do it but if you have one you need to create no drop qos and merge it in |
| Storage VDC configuration | (1)interface in same pair-2 consecutive 10G configured(2)trunk mode(3)shut in parent VDC(4)vdc fcoe id 2 type storage(5)(config-vdc)allocate int e4/1-16,e4/19-22(5)vdc fcoe id 2(6)(config-vdc)allocate fcoe-vlan-range 10-100 from vdcs N7K(7)allocate shared int e4/17-18 |
| Storage VDC config #2 | (1)enable lldp(2)enable lacp(3)enable npiv |
| vlan to vsan mappings | (1)vsan database(2)(config-vsn-db)vsan 50(3) vlan 50(4)fcoe vsan 50(5)config LACP |
| LACP load balance best practice | (1)port-channel load-balance ethernet source-dest-port - is best for FC for 5K |
| FCoE npv vfc port creation | (1)interface vfc-port-channel 50(2)switchport mode f(3)switchport trunk allowed vsan 50 - 51 - NOTE: vfc automatically bound to Ethernet port-channel with same number for MDS and N7K |
| FCoE target config | (1)int e4/17(2)switchport trunk allowed vlan 50(3)interface vfc417(4)bind interface e4/17(5)switchport mode f(6)switchport trunk allowed vsan 50(7)vsan database(8)(config-vsan-db)vsan 50 int vfc417(9)no shut |
| FCoE initiator on 5K | (1)int e1/9(2)switchport mode trunk(3)switchport trunk allowed vlan 50(4)spanning-tree port type edge trunk |
| VFC interface config for VN CNA on 5K | (1)int vfc119(2)switchport mode f(3)bind ethernet 1/19(4)switchport trunk allowed vsan 50(5)vsan database(6)vsan database(7)vsan 50 int vfc119(8)no shut |
| FCOE VE configuration | nothing changes except use an E instead of an F |
| EvPC & FEX | best practice is a FEX straight thru |
| can do Enhanced vPC | (5KA config)(1)fex 100(2)(config-fex)fcoe(3)fex 101(5KB)(1)fex 101(2)(config-fex)fcoe(3)fex 100 |
| FIP process | (1)FCOE vlan discovery - FIP multicast to ALL_FCF_MAC look for FCOE VLAN FIP frames use native vlan(2)FCF discovery find FCFs answering for FCoE vlan - FCF's respond back w/their MAC(3)Fabric Login-FIP sends FLOGI to FCF_MAC from step 2 - establish virtual link host & FCF |
| FIP troubleshooting | (1)show lldp int e1/4(2)show system internal dcbx info int e1/2 (should see DCX = CEE) - look for errors - check dcbx counters (3)show dcbx counters(4)get packet capture(5)show platform software fcoe_mgr info int vfc18 |
| OTV benefits | (1)MAC-in-IP over any transport(2)protocol learning(3)preserve failure boundry(4)built-in loop prevention(5)automated multi-home(6)site independence(7)dynamic encaps(8)no psuedo-wire maintain(9)optimal multicast replication(10)multipoint connectivity(11)point to cloud model |
| Edge Devices | (1)perform all OTV functionality(2)located at aggregation or core layer(3)support for multiple edge devices aka multihoming |
| internal Interface | (1)site facing interfaces of the edge devices(2)carry VLANs extended through OTV(3)regular L2 interfaces(4)no OTV config required |
| Join interface | (1) one of the uplink of the edge device(2)PTP routed interface(physical int, sub-int, or port-channel)(3)used to physically join the overlay network(4)no OTV specific config required(source or destination ID) |
| overlay interface | (1)virtual interface with most of the OTV config(2)logical multi-access multicast-capable interface(3)encapsulates L2 frames in IP unicast or multicast |
| AED | Authoritative Edge Device(1)OTV suppports multiple edge devices per site(2)single OTV device is elected as AED on a per-VLAN basis(3)the AED is responsible for advertising MAC reachability and forwarding traffic into and out of the site for its VLANs(4)is a loop prevention mechanism |
| AED election | there is an election to see if they are the authoritative edge device for even or odd vlans |
| Site VLAN | L2 vlan not extended across overlay just to exchange hellos for edge devices |
| Site Identifier | verify that they are members of a site based on an identifier that is layer 3 capable |
| Site Adjancency | established across site-vlan - DO NOT EXTEND OVER THE OVERLAY |
| overlay adjacency | established via join interface across L3 network |
| Communication problems | even if there is one on the site vlan, each OTV device can still advertise AED capabilities to prevent active/active scenario - if is down will become AED for all VLANs |
| config site vlan and site identifier | (1)otv site-vlan 210 (2)otv site-identifier 0001.0001.0001 |
| OTV MTU | (1)otv adds 42 bytes header to both ISIS control traffic and encapsulated data traffic(2)OTV sets DF bit on all packets(3)transport must support 1442 bytes for multicast and 1450 for unicast(4)make sure transport supports at least 1542 though for most apps-jumbo frame enable |
| OTV MAC Table | (1)no unknown unicast flooding(selective post 6.2)(2)control plane learning with proactive MAC advertisement(3)background process with no specific config(4)IS-IS used between OTV edge devices(5)if not known MAC it drops it |
| OTV and MAC advertisement | in order to advertise MAC OTV edge device must do these (1)discover each other (2) build a neighbor relationship |
| Neighbor relationship | (1) multicast ernabled (2)unicast - post 5.2 |
| OTV config | (1)feature otv(2)int overlay 1(3)int e1/9(4)(config-if)ip igmp v3(4)otv join-interface e1/9(5)otv control-group 239.1.1.1(6)otv data-group 232.1.1.0/24(7)otv extend-vlan 110(8)otv site-vlan 200 |
| adjacency server | enables you to provide unicast-only transport between edge devices when ip multicast not available - configure a primary and secondary |
| show otv adjacency | shows otv adjacency and all its peers |
| show otv route | shows mac routing table |
| show otv overlay [if-name] | displays OTV overlay interface config |
| fabric merge guidelines | (1)unique domain IDs(2)zone names unique but if not must contain same members(3)zone mode basic or enhanced must be same - show zone policy vsan x |
