This is the Q&A from "Cisco Nexus: Overview and Architecture".
Nexus Overview and Architecture Related:
Q: to commit configurations do we need to use commands like "commit' as in IOS XR or do a wr mem in IOS /IOS XE ? A: We use Copy run start in the Nexus
Q: when you say 8 VDC's does it include the default VDC ? A: Yes it includes the default VDC. These combine to allow some significant increases in scalability across the chassis, primarily with the number of VDCs. Supervisor 2 supports 4 VDCs while Supervisor 2E supports 8! Additionally a new capability called “Admin VDC” comes in NX-OS 6.1(1) on the Supervisor 2/2E so you’ll frequently see the VDC count listed as 4+1 and 8+1 with the +1 being the Admin VDC. More details on Admin VDC in a bit.
Q: Is it possible to connect storages directly to nexus using a FC interface? Or do we need a Fiber Interconnect which will convert the FC to FCoE? A: you can connect FC storage directly to N5ks
Q: why we have different OS now, NX, XR and XE? A: NX-OS operates on Monta Vista Linux, Kernel only. and IOS-XR is the next gen, we are moving away from IOS for its rich robustness
Q: kindly explain VDC,FEX... A: VDC is a virtual switch in physical chassis. Lets say you have one Nexus physical chassis and you can create 4 virtual switch with Sup1 and 8 virtual switch with Sup2. FEX is just a module extension where you can connect more devices.
Q: Single Sup only 4 VDC but what if we need to increase VDC..do we buy license or a new SUP A: when we say SUP-1 its not one single supervisor its not the count, its the SUP-1 generation like we have in 6500 sup32, sup720. Sup-1 and Sup-2 are the only N7k supervisor engines available now and both needs different licenses as license is restricted to chassis
Q: Is Nexus N2K and 5K different from N7K A: yes N2k is a FEX which is just a dump box does some forwarding and requires a parent switch like N5k or N7k. N5k and N7k are different architechtures although few of the NXOS architechture remains same
Q: what is the command to read the debug on the switch itself ? A: Enable, Debug logfile <name> and then enable the debug ip route for example, then you will see a file generated in the logflash called <name>.you can then copy the file from the logflash to bootflash then do show file bootflash:name for reading it in CLI
Q: EIGRP and OSPF are handle by Pak_priority in nexus? A: by default we comply with the RFC 791 and RFC 2474. So yes, there is PAK-Priority in N7k. Since we have CoPP and Rate-limiter much more enhanced than IOS we will not face such situations like IOS
Q: which is the commad to see the captured debug outputs in switch itself A: Enable, Debug logfile <name> and then enable the debug ip route for example, then you will see a file generated in the logflash called <name>.you can then copy the file from the logflash to bootflash then do show file bootflash:name for reading it in CLI
Q: Process reload = how is this achieved? A: If there is a software exception unlike IOS, we just reload the process, this is done in the kernel level and Core will be stored in the logflash. This is completely transparent to the end user
Q: what is meaning of parellel upgrade A: Parallel upgrade is for upgrading 2 or 3 I/O modules upgrade at one single point of time when using ISSU
Q: question was on commiting the configuration in NX-OS A: We dont have Commit Option in NXOS and we have COmmit only for few features and not for all like ACL
Q: when reloaded the process - could this have an effect on related config? ie reload / restart hsrp process could this remove the hsrp config current applied? A: Reload process is a stateful one, Configs is stored in the PSS. we wont remove the configs instead we just restart the process
Q: what is maximum mtu supported on nexus devices A: Jumbo MTU 9216
Q: if my Nexus 7010 chassis is loaded with 2 Power supply and and taking a load of 8000W, we are using 6000x2 PS A: Please use power calculator @ http://tools.cisco.com/cpc/ If one of your PS goes down the other PS takes on and work at 100% of its capacity
Q: maximum vdc in nexus 7010 A: VDC count is not related to Chassis whereas its related to SUP-1 and Sup-2. Sup-1 supports 4 VDC and sup-2 and SUp2E supports 8 VDC
Q: then hw the nexus will behave A: If one of your PS goes down the other PS takes on and work at 100% of its capacity
Q: does the communication between two ports on two different line cards , will go to SUP engine for processing A: No packets will go to SUP except the control-plane packets. Flow through will from LC1---XBAR---LC2 no sup in the path
Q: How to troubleshoot hardware issues related to fabric module? A: We have extensive procedure for troubleshooting XBAR modules which we wil cover in Session - 2
Q: whether it will power down the whole chasis or it will power down some modules , if yes which modules in wat order ? A: The default behavior of the EEM script is to power down the I/O Modules in the order of 1-4, 7-10 (missing the Supervisors).
Q: What is the over subcriptin ration on these modles A: Each modules have different Oversubscription ratio and F-series are linerate modules and m series where these subscription comes in. Please search datasheet for respective modules for more details
Q: how would you differentiate cisco 6509/6513 and 7010/7018 A: Both are different chassis one is for 6500 and other is for Nexus 7k. If you are referring sup architechture, both are completely different
Q: what is the over subcription ratio on these models? A: Each modules have different Oversubscription ratio and F-series are linerate modules and m series where these subscription comes in. Please search datasheet for respective modules for more details
Q: Why did Cisco do away with the CMP on the newer SUPs? It seems like a nice feature A: Most of the customer base doesnt use CMP at all and that has lot of bugs although its a seperate COmplex
Q: Hi, Do u hav any documentation where we can block the HSRP hellos on the Four N7K which are particpating in Fabricpath. So two N7K devices form one bundle & HSRP remins Active/Stanby state on one bundle only for same vlan which is on all the four N7K? A: Please use different HSRP version thats the easiest way to block HSRP hello;s or use HSRP authentication key.
Q: plz share me the difference betwnn VSS and VDC? A: VSS is a 6500 bundle with one active control plane and one standby control plane whereas the VDC is a virtual switch in one single physical chassis
Q: what is the maximum mtu does the nexus port support and how to change A: In N7k, we use mtu command before 6.2.2a and after 6.2.2a we use QoS polcies to change the MTU, Max MTU is 9216
Q: sup2 and sup2E both support only 2.13 GHz while sup2E has 2 quad core A: The Supervisor 2 module (N7K-SUP2) has a quad core CPU and 12 GB of RAM. The Supervisor 2 Enhanced (2E) (N7K-SUP2E) module has a dual quad core CPU and 32 GB of RAM and provides increased software scale such as an increased number of VDCs and con
Q: what is the main difference between M and F modules ? A: M and F are different family series, F series has a Switch on Chip architechture where every port functions at line rate and gives you additional features like FP. M-series is a high PPS card has blocking and non-blocking cards.
Q: can we mix the LC between F and M series to build vPC A: while building VPC we need same module type
Q: i have the only F series card can i create the L3 vlan interface on it A: F1 is a layer2 card, F2 you can create SVI.
Q: w.r.t my HSRP question on fabricpath ....the HSRP version, HSRP group & Vlan number remains same on all the N7K....tht is requirement....so can u share any documentation on same. thanks. A: ok, if thats the requirement, you can use PACL in the ports where the hellos are sent out and block just the hellos
Q: can you share link which compares all 2k, 5k and 7K in single sheet or PDF with all details? A: No such document exists, if you please be specific i can answer you questions on the difference
Q: Cisco 6509 support 2 SUP module having SSO and RP(Route Processor) failover. A: yes we have the same here, two SUP and in HA which is called high availability not SSO as in IOS
Q: how many dedicated port can be configured with M1 card A: Please be specific on which M1, M132, M148, M108 which one?
Q: main purpose of SOC plz A: Switch on Chip (SoC) does everything right from the layer 3 forwarding, replication of the packet, MAC address learning etc
Q: how do u configure particular vlan on a specific SOC ....is ther any specific command A: Please dont get confused, SoC is internal to F-series for understanding purpose, in F1 - each 2 ports goes to one Soc and in F2 - each 4 ports goes to one SoC.
Q: Is it possible to knock down the 40G ports into 4 x 10G SFP+ interfaces? if yes, does the 10G interfaces support normal link distances like upto 10 km for single mode OFC? A: yes we call that as breaking the modules. There are limitation with respect to distance please be in touch with sales team
Q: thts correct...as there was a comment by Sudhakar for scalability of MAC address ...we need t oconfigure specific vlan on specific SOC to scale to 196K mac's A: yes, so you need to allow VLAN's only to a specific SOC. IF you allow vlan like vlan 2 in soc 1 and soc 2 then both should be in sync
Q: it was bit confusing answer on SOC you mentioned.....if ther eis any doc which epxlain in detail that would be helpful. Thanks A: Please refer NXOS architechture from cisco Live presenter: Tim stevenson BRKARC-3470
Q: All this IOS has QoS capability ? A: NXOS has QOS capability differs with Linecard and only MQC support is there
Q: VoQ is valid for Fab2 as well?? A: Yes VOQ(Virtual output queueing) is a architechture and is applies to Fab2 as well
Q: will there be any issue with VPC when running in between 2 NX7k having diff NXOS ver 6.x A: Running different NXOS between the different VPC peer-link is supported however its not recommended to run them with different version
Q: Does N7K suporrt FCOE? A: you can create SAN VDC and yes it supports FCOE
Q: is it TACAS intigration is deficult in NX-OS A: Its easy compared to other boxes, we have DCNM which tightly integerates with the Nexus platform and does it smoothly
Q: Are All NExus are MLS ? A: MLS is a concept in 6500 and we dont use MLS on Nexus its different architehcture
Q: Please elaborate on Fabric Path A: Fabricpath is a trill based technology and we will cover that in the second session
Q: so i need to dedicate a VDC for SAN.....and it doesn't support FCOE on the default VDC A: yes you need a dedicated VDC thats called SAN VDC
Q: Is NX-OS are same for all series switches? A: NXOS train's are different for N5k and N7k and N3k. N2k download the image from the parent platform i.e, N5k,N6k or N7k
Q: will VPC works in between SUP1 & SUP2E , having same NXOS ver A: yes it will work however its not recommended
Q: Please paste the link for comparison between NX-OS and IOS A: There isnt any link which shows immediate comparison however you can compare them feature wise by reading whitepapers for both the platforms
Q: Can all Nexus switches will be used at all envirnmnet or its only for Data Centre specific.? A: N7k is built for Datacenter and yes you can use it in all environments
Q: can we use dirrerent version of kickstart and system image together A: no you cannot use it, system and kickstart has to be in the same version
Q: commad write ersase does not work on Nexus ? A: write erase works with nexus
Q: can we connect nexus to cisco 6500 ? is nexus compatible with cisco 6500 ? A: yes it is compatible with any platforms