Ask the Expert: Nexus 5000, 3000 and 2000 Series

[ciscomoderator]

With Lucien Avramov

Welcome to the Cisco Support Community Ask the Expert conversation. This is an opportunity to learn about design, configuration and troubleshooting of the Cisco Nexus 2000, 3000, 5000 Series with cisco Expert Lucien Avramov. Lucien Avramov is a technical marketing engineer in the Server Access Virtualization Business Unit at Cisco, where he supports the Cisco Nexus 5000, 3000 and 2000 Series. He has several industry certifications including CCIE #19945 in Routing and Switching, CCDP, DCNIS, and VCP #66183

Remember to use the rating system to let Lucien know if you have received an adequate response. 

Lucien might not be able to answer each question due to the volume expected during this event. Remember that you can continue the conversation on the Data Center sub-community   discussion forum shortly after the event.   This event lasts through October 21, 2011. Visit this forum often to view responses to your questions and the questions of other community members.

[Lucien Avramov]

Correct that is in case you have a failure of N5K. If you have a fabric extender that is dual homed to two nexus 5000 switches, then this will allow you to have active / active

[areanetsenato]

Hi Lucien,

I've some doubts about stability and reliability of Nexus 5000 firmware...

4.2(1)N2(1)  Release Date: 30/JUL/2010

5.0(2)N1(1)  Release Date: 18/OCT/2010

5.0(2)N2(1)  Release Date: 20/DEC/2010

5.0(3)N1(1)  Release Date: 03/MAR/2011 --> deferred

5.0(3)N1(1c) Release Date: 07/MAY/2011

5.0(3)N2(1)  Release Date: 13/JUN/2011

5.0(3)N2(2)  Release Date: 09/SEP/2011

After 4 update in 5 month and after a deferred version i've stopped upgrade on 5.0(3)N2(1) because i can stop my network so often...

Nexus firmware 5.0(3)N1(1c) automatically reboot our two Nexus 5010 every 12/15 days; support suggestion was  update to 5.0(3)N2(1) and the new version was up for 28 days... a great improvement !!!

And then I've read incomplete Release Notes or caveat resolved but not listed...

What is happening ?!?

I've installed 4 Nexus 5010 and 12 N2K-C2248TP and every firmware upgrade is a conflict with my management due to real cause and resolution for the new firmware update ...

I cannot continue with this level of service...

Any suggestion ?!?

Regards

Marco

[Lucien Avramov]

I'd suggest to work with our TAC engineers and escalation team to exactly find the issues you were running into and you can have them also involve me so I can work with you on a get well plan. There were reasons for the deferral and we had introduced new platforms and new software features which required to have a few releases in 5.0.3 code. In general due to the high feature velocity, the latest version is the greatest, here for instance 5.0(3)N2(2), but I would suggest to first identify precisely the root cause of your current issue(s).

[mleinio]

Hi Lucien! What is the maximum number of hardware port-channels in Nexus 5548 and 5596? If I remember correctly it was 16 in 5010 and 5020. Is the hw port-channel allocation logic also the same as in 5010 and 5020?

[Lucien Avramov]

I'm glad you asked, it's a higher number on the 5500 series generation of switches.

There is no limitation besides the number of ports here.

You can have up to 48 port-channels on 5548 and 96 on the 5596.

We have a document that shows limits on both platforms here, it's called configuration limits:

http://www.cisco.com/en/US/products/ps9670/products_installation_and_configuration_guides_list.html

[GarryGlendown]

Hi,

we have just put a new core switch up for a customer, currently one 5548 feeding L2 into several 2960S switches, as well as 4 FEX 2248 (with 2x10GE each) for Data Center access in two locations. The 5548 and two 2248 are in one data center, the other two 2248 are in the second DC. Early next year the customer will be adding a second 5548 for full redundancy (currently, the servers in the DCs are connected to both 2248's in DC, which doesn't help if the 5548 crashes or dies). I've seen docs about doing a dual 5548 configuration with the FEXes connected to two N5K's. What do I need to watch out for when extending the current configuration? How is the configuration of the N5k's handled in respect of the FEX ports? Manually I assume?

[Lucien Avramov]

When you will extend the configuration, you will be able to dual home the FEX to the Nexus 5000s, each FEX being dual connected to both upstream Nexus 5000. This will be what we call a vPC design. Depending on the number of ports free on your 2960 / 2248s, you could first configure both Nexus 5000s in vPC, then one FEX on each side dual home in vPC at the time. Migrate the server ports to the other FEX or on the 2960, then when the FEX comes back online in ''vpc'' topology you can move servers to this FEX and upgrade the other one in each site. This would minimize the down time in the design topology change.

Here is a document about quick start on vPC, a design guide and a operations guide:

http://www.cisco.com/en/US/prod/collateral/switches/ps9441/ps9670/configuration_guide_c07-543563.html

http://www.cisco.com/en/US/prod/collateral/switches/ps9441/ps9670/C07-572834-00_STDG_NX-OS_vPC_DG.pdf

http://www.cisco.com/en/US/docs/switches/datacenter/nexus5000/sw/operations/n5k_vpc_ops.html

[ike1cisco]

Hello I have a question about vPCs ..relating to being visble to the third switch via port-channel.Correct me if am wrong and expand on this staement please " Allows a single device to use port channel across two upstream devices'. Am trying to grasp the main comcept of vPCs and how to verify them in a live environment on the involved devices assuming its spannoing through more than two Nexus switches

[Lucien Avramov]

vPC always works for a pair of Nexus switches (7ks, 5ks, 3ks). The pair of switches will be configured as vPC and will then appear as a ''single'' logical switch from the point of view of the device you port-channel to them (one or more link to each).

Here is more information about the vPC design:

http://www.cisco.com/en/US/prod/collateral/switches/ps9441/ps9670/C07-572834-00_STDG_NX-OS_vPC_DG.pdf

Some documentation about how it's configured can clarify the concept:

http://www.cisco.com/en/US/prod/collateral/switches/ps9441/ps9670/configuration_guide_c07-543563.html

Here is also a webcast about 5k and vPC (the concept applies as well to N7k and N3k):

https://supportforums.cisco.com/videos/1228

[smartboy2255]

HI Lucien,

   I have one problem , I have two N5k with single Nortel Core switch.My uplink is 1G ports.I was able to ping the outside the network.But when access data ,it wont work. If I remove the redundant link it is working fine.Now I have upgraded firmware from 5.0(2) to 5.0(3). Now everything is working fine.What could be the problem.Can you pls explain ?

Regards,

Ajith

[Lucien Avramov]

I'm guessing you have Nexus 5548. On the 5500 series, we started supporting 1G speed from 5.0.3 version.

[cnrs-dsi]

Hello Lucien

We will buy a pair of Nexus 5000 to interconnect racks containing  Dell chassis. They each contain a pair of switches 3130X. In a rack  there are four chassis.

We would like to make a stack of 3130X inside the rack and connect it with the pair of Nexus using VPC.

The goal is to have a switch at the logical level (the stacks of 8  3130X) connected to a swicth (the pair of Nexus). These two technologies  can be combined?

Thanks

[Lucien Avramov]

As long as it's compliant with dot1q and lacp there should be no problem.

[tiwang]

hi lucien

even though it is not a vpc forum I'll try this question - how do I verify load-balancing between the nexus boxes in a vpc domain? I have a set of nx5020 which holds a vpc domain - this vpc is interconnected to another set of nexus boxes which also holds another vpc domain. These two vpc domains are interconnected with  2 10 gig links - and I hoped that I somewhere/somehow could verify load-balancing and see how many links participate in this vpc etc - but hmm - ? - well as far as I can see it is a bit cumbersome to verify which links participate into a vpc and the traffic- the boxes exchange a lot of info about the state - are there some commands here I don't know about or something on the road map?

best regards /ti

[Lucien Avramov]

Ti, from a server or a switch attached in a vPC to both Nexus switches, this is then the principle of hashing: the device will hash to either one of it's link going to the Nexus switches. When a packet comes through the Nexus Switches going to this end device depending on which one received the packet it will send it on that link.

You can see the interface bw usage via the regular show commands.

Please take a look at the vPC section on this webcast:

https://supportforums.cisco.com/videos/1228