UCS < 1.4.2b >, N1K and uplink

Similar Questions

• Channel ports (connection linked to the North) and uplink interface (private storage switch)

  Hello

  What will be the best configuration if I connect 2ports north bound connection and has been applied to the port channel. and I go to configuration 2ports uplink for storage (private switch) interface. All cisco switches.

  2 x 2104XP IOM

  2 x FI 6120XP

  Thank you.

  Dexter

  Hi Dexter,

  Since you will have two / several switches uplink to which you will have an Ethernet to the North, a what you will have to configure is the layer 2 disjoint.

  You must be on any version 2.0 to configure that.

  Here are a few documents to help you with the details:

  http://www.Cisco.com/en/us/solutions/collateral/ns340/ns517/ns224/ns944/white_paper_c11-692008.html

  http://www.Cisco.com/en/us/docs/unified_computing/UCS/SW/GUI/config/Guide/2.0/b_UCSM_GUI_Configuration_Guide_2_0_chapter_010101.html

  Thank you!

  . / Afonso

• Cisco 2960 and uplinks AIR-AP1041N

  Guys,

  what I'm doing wrong here?  I need to enable trunking to pass on several VLANs to my AP, but when I put the 2960 as trunk mode I lose connectivity to the AP.

  All will help it useful!

  Course AP Interface congfigs:

  interface GigabitEthernet0.210
  210 native encapsulation dot1Q
  no ip route cache
  Bridge-Group 1
  No source of bridge-Group 1-learning
  Bridge-Group 1 covering-disabled people
  end

  +++++++++++++++++++++++++++

  210 native encapsulation dot1Q
  no ip route cache
  Bridge-Group 1
  Bridge-group subscriber-loop-control 1
  Bridge-Group 1 block-unknown-source
  No source of bridge-Group 1-learning
  unicast bridge-Group 1-floods
  Bridge-Group 1 covering-disabled people
  end

  2960 new switch port config (lost connectivity to AP unless I put back to access the mode vlan210 why?)

  GigabitEthernet1/0/1 interface
  AP USTCRLW001 description
  switchport mode trunk
  SRR-queue bandwidth share 1 30 35 5
  priority queue
  MLS qos trust cos
  Auto qos trust
  end

  Add this on swicthport where AP is connected:

  GigabitEthernet1/0/1 interface
  AP USTCRLW001 description

  Switchport tunk native valn 210
  switchport mode trunk
  SRR-queue bandwidth share 1 30 35 5
  priority queue
  MLS qos trust cos
  Auto qos trust
  end

  Concerning

  Remember messages useful rates

• FEX Uplink and Ports on the side Server

  I am very new to the environment of the UCS and having to return to the current fairly quickly. I decided to start with the chassis. But have a few questions.

  I hope someone can clarify how serverside ports are assigned less than ports uplink, 2204XP and 2208XP.

  My understanding is that the side ports server are assigned to the uplinks based on the number of uplink configured at the END.

  Once the FEX is initialized ports are hardcoded.

  Assuming 2 (6200 series) FI and 2 FEX (2204).

  2 each to their respective FI FEX uplinks

  8 blades width in the chassis.

  My understanding is;

  Uplink port 1 receive blades (1, 3, 5-7) and uplink 2 (2, 4, 6, 8)

  Question: Which ports are active for each blade and that FEX.

  For example the blade 1 assigned to Uplink on FEX A 1 with FEX B 1 uplink failover

  Blade 3 assigned to Uplink on FEX B 1 with FEX A 1 uplink failover

  Thanks in advance people...

  Hello and welcome to the community,

  "" My understanding is that the side ports server are assigned to the uplinks based on the number of uplink configured at the END.".

  Server ports are assigned to connections rising from ports either automatic (load balanced by available uplink interfaces) or manual pin (PIN LAN groups hard coded)

  2 fIs and IOM 2204 2 with 2 cables on each IOM: "UpLink port 1 receive blades (1, 3, 5-7) and uplink 2 (2, 4, 6, 8)" "< this="" is="" correct,="" odd="" blades="" through="" odd="" ports="" and="" even="" blades="" through="" even="">

  Which ports are active for each blade and that FEX.?

  In this case with a 2204, IOM provides 16 internal ports, 2 for each blade, and you can check which one is active with the command "Show software platform woodside m" mode FEX which is realized with the following commands:

  UCS # connect the IOM

  FEX-1 # display software platform woodside m

  

  You can also go to the vNIC itself in the "Servers" tab by going to the Service profile > network and then double click on in the vNIC and you will see the configuration:

  

  For example: blade 1 assigned to Uplink on FEX A 1 with FEX B 1 uplink failover

  Blade 3 assigned to Uplink on FEX B 1 with FEX A 1 uplink failover

  This will be the case if you have the 'discovery of the chassis/FEX policy' located in the 'grouping preference link' on 'None' instead of 'channel of Port "; If it is "Port Channel", I won't miss not until all other links fail in the same module e/s, see below

  

  NOTE: this option is valid only if your hardware is ALL second generation, in other words, it does not work with fabric 61xx interconnections or 2104 IOM

  I hope this helps, if it does, please note it.

  -Kenny

  Note ensure on the limitation of the function of the channel port between IOM and the tissue Message was edited by: Keny Perez

• SG500X - stacking and 10GbE Uplink together

  Hi guys,.

  I use a SG500x in native mode of stacking, using 10g ports S1/S2.

  My question is, in this mode am I could simultaneously use 10 GB uplinks/XG1 XG2 ports? The manual is not clear at all!

  If I can both stack and uplink at the same time can someone also let me know if I get the full 40 GB of bandwidth?

  Finally I need a SPF 10 GbE module so I can run cat6 from the ports of X 1 / X 2 for 10 GbE on my cluster but network cards can work on which we're compatible!

  Sorry lots of questions!

  Thank you guys

  Hi Sam, you can use the ports of the battery and the other 2 10 gig. S1/S2 are the default stack ports. The S3/S4 would be considered ports network at this time here, but can be toggled to use S3/S4 instead of S1/S2.

  The ability of the switch is 128 Gbps on the model of port 24 th, 176 Gbps on the 48-Port models. So, if you say a 48 switch ports.  1 = 48 x 48. 10 x 4 = 40. 40 plus 48 is 88.  88 x 2 = 176.

  You should get the result you want (if you can actually use the link 10 gig to capacity, many Internet tele-conferences are not yet used to full 10 gig).

  For the SFP capatibility, you can reference the Administrator's guide. It has all the modules support. Once you choose your module support you should talk to your provider of NETWORK card with one should work with your cluster.

  -Tom
  Please mark replied messages useful

• Problem of VM - FEX - uplinks in VmWare are blocked by admin

  Hello

  Im trying to configure the VM - FEX on our infrastructure of UCS. I've updated all the components of the UCS for version 2.1(3a). Our VmWare is verision 5.1.0 and we have Enterprise Plus feature set.

  I have studied several documentations on the VM - FEX configuration, but without success...

  I have a UCS VCenter integration upward. My DVS created in the UCSM can be seen in the vCenter, also when I create the port profile, I see in the vcenter, and attach the virtual machine with it.

  The only problem is, the uplinks to the VM - FEX DVS are in the State, "blocked by the admin" in the vCenter, and there is no possibility of silent them.

  Can someone help me?

  Thank you and best regards Jiri.

  Initially, we used to have the same file for the N1k and VM - FEX but this sometimes changed back. A good way to check this would be to see what VEM modules are loaded. For example:

  N1k:

  ~ # esxcfg-module -l | grep vem vem-v160-l2device        3    36 vem-v160-n1kv            13   220 vem-v160-vssnet          4    6820 vem-v160-stun            1    128

  VM - FEX:

  ~ # esxcfg-module -l | grep vem vem-v152-vmfex-svs-mux   11   32 vem-v152-vmfex-pts       1    148

  (V160, v152) version number could vary, but the modules PTS, svs - mux etc will not change.

  Thank you

  Shankar

• UCS mac address problem

  Hi all

  simple question...

  You may know if it is a bug of mac address learning with version 1.4.1?

  I have go a blade with a mix of palo and other cards, some vnic are pinned to link rising with trunk allowed vlan 501 on both sides. On the blades of the UCS, i've got ESX and ESXi. Service console is labelled VLAN 501 (depending on the configuration of the trunk switch UCS and uplink). Some blade had obtained these interface pinned on the UCS6100 A and another on B (with failover). The fact is that if I ping from outside UCS UCS (ESX) inside no learning of mac is performed on all infrastructure.

  I don't know problem is the UCS because if I erase the mac table UCS, nothing works... and all the works if I start doing a ping of the UCS. I also tried to put the switch uplink static mac (ESX service console) but was not helpful.

  I hope this isn't a 1.4.1 bug...

  TNX and tnx for help

  Dan

  Y at - it you using uplink groups pine or have a network of layer 2 disjoint by chance?

  Questions like these are ususally PIN-related groups / Disjoint L2 using the wrong link DR port for traffic entering.  When the traffic is generated from within the outgoing UCS, the MAC table is updated on the upstream switch which allows traffic to flow, until he gets older.

  If one of them is the case I suggest take you a look at the videos of Brad name video #6:

  http://bradhedlund.com/2010/06/22/Cisco-UCS-networking-best-practices/

  Kind regards

  Robert

• Problem with MAC pinning and new VLAN

  [Cross Announces nexus 1000V and forums of the UCS]

  Hi all, I have a working of the UCS 1.4 configuration (3i) and the Nexus 1000V 1.4 and according to best practices guide uses "channel-group auto on mac - pinning mode" on the uplink of Nexus. I have a problem when you add a new VLAN in this environment, and it is a reproducible problem in two different facilities.

  I pass by the VLAN usual creation process on the Nexus, the network upstream and within the UCS himself. I create new port vethernet profile and set it as an access port in the new VIRTUAL LAN. However when I join a VM (existing or new) to this new port vethernet profile in vCentre virtual machine cannot communicate with anything. However, if I disable MAC pinning with 'no channel-group auto on mac - pinning mode', the virtual machine will start instantly talk to the outside world and the new VIRTUAL LAN is running. Can I turn MAC pinning back once again, and everything continues to work.

  So the question is, is this normal or is there a problem? MAC the Θtiquetage has a brief interruption of the uplink, so is not a viable long-term customer solution when they want to add new VLAN. Is it possible to add new VLANS in this scenario without any unavailability of the network, but brief?

  Thank you

  Close the loop on this.  You tap the CSCto00715bug.

  Symptom: New MAC address is not learn on vem in the l2 table even though the mac address table is not overflow yet. vemcmd show l2-emergency-aging-stats | grep "Number of entries that could not be inserted:" will show extreme large number. Conditions: Nexus1000v VEM running on SV1.4 release. There are two CPU cores on the host. This issue may happen at race condition. Workaround: Reboot the ESX/ESXi host. 

  This problem is corrected in the 1. 4A release.

  Kind regards

  Robert

• Effect of allowing disjoint networks UCS L2 for existing pod

  Servers within our UCS pod existing v2.0 (3 c) need new connection to a network of layer 2 in our datacenter. This network is not accessible to the Juniper switches to the North of our current 6248 fabric of interconnections.

  After reading the documentation, it seems that "disjoint networks layer 2" may be the answer. If I understand correctly, we would have cable redundant connections to 6248 unused ports to switches that host this network of layer 2 and follow the guide associated with that vlan L2 these ports.

  However, I also understand that I need to tweak the configs for our existing fabric and the fabric port channels B that hosts all the VLAN "normal." I fear that doing so may impact on the network for computers UCS hosts on these networks, or may necessitate a restart/reacknowledgement of existing hosts of UCS. Am I being too pessimistic?

  See you soon,.

  Paul

  Hello Paul,

  I would indeed provide a window of failure because you configure VIRTUAL LAN Manager to associate some VLANs for the uplinks on each FI, it will be re-pinn all vEths which will cause a slight disruption in traffic.  According to the degree of sensitivity are your applications they may or may not be affected - I'd do it wrong side of caution and take a maint 1 hour. window to set up and test accordingly.

  By default, all them VLANS are allowed on all uplinks.  As soon as you add a VLAN to an uplink channel / port specifically via the VIRTUAL LAN Manager, he gets explicitly cut off the coast of all other channels of uplinks / port for FI.

  A great way to check your configuration is once you have everything configured as you see, you vous connecter connect to the UCS CLI interface, "connect nxos" and run 'show int trunk '.  You should see the VLAN on uplinks necessary - with the exception of the VLAN 1, that will be allowed on all uplinks.

  Kind regards

  Robert

• EHM: 1 uplink port failure

  Hello

  For now, to start a new area of UCS, my client would like to set up an Ethernet Uplink with 1 single port/FO. So, we set up a purchase order with 1 port.

  If this link fails, do all the vNIC using this uplink fail?

  I mean, the WMare vswitch vmnic are down so this Ethernet uplink failure?

  Thank you.

  Nicolas.

  Hi Nicolas,.

  If you have enabled vNIC failover UCS level, the vNIC would move to other interconnection fabric. For example, if the vNIC initially points to FI and the port-channel on FI breaks down, with resume vNIC UCS function would kick in and move the FI B connection. The vmnic on ESX will not be aware of this.

  If you have not enabled failover, then the NIC would go down on the UCS (and ESX)

  Thank you

  Shankar

• Half of the blades fell when creating a new channel of uplink port?

  I have a pair of 6248UP fiber interconnection (A and B) with blade servers connected. The switch upstream is Nexus7k. Fabric A and B each right has now a channel of uplink port to the Nexus switch upstream. I add a new Nexus switch to connect to the interconnection of fiber. The plan is to have uplink some VIRTUAL networks for Nexus one pass and uplink some VIRTUAL to the second Nexus switch networks.

  After that I configured the new Nexus switch with a channel port (only allowing some VIRTUAL networks) and corresponding port created on fabric channel A and B, new ports-channels have increased. However, half of the blade servers were down after that. What could be the reasons for this problem? Is it because half of the blades have been pinned dynamically on the new channel port, but the new port-channel side Nexus allow only some VIRTUAL networks?

  Is there a way to configure the Channel ports on fiber to interconnect with only the trunk through some VIRTUAL networks? for traffic from the server would not be sent to the port-channel where VLAN server is not allowed?

  Thank you

  In general, all NICs for/vhba of a fabric are declining, if all uplinks of this fabric are declining.

  It's not clear to me, what you're trying to achieve?

  If you create a global VLAN (for the two fabrics), this vlan is automatically inserted in all dependent North trunk links, unless you configure a leaking installation

  (http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ns224/ns944/white_paper_c11-692008.html)

  Create a new vlan on UCS requires of course also this vlan is created on the switch connection and vpc by default allows all the VLANS. Could it be, that the vlan not appeared on the switch and was therefore not upward on the trunk?

• FEX-Uplink Failover, loadbalancing

  I would like to understand more of the technology within a vblock
  I have a few questions gradually

  So I'm out. pls be patient and thank you in advance...

  VBlock 720 with the following generic configuration

  Switch 2 x Cisco 5000 Series
  2 x FI 6248
  6 x chassis 5108
  FEX 2204
  VIC 1240 (no Mezzanine Card)
  8 x M3 blade, 2 CPU

  uplinks of 4 x 10 G of each FEX to FI (port channel)
  uplink of 8 x 10 G of each FI for Ciso 5000 series swith (vpc)

  Assumptions:

  All chassis are completely fill

  Each blade has consistient thoroughput of bandwidth, but 6 GB

  1. When the blades connect to FEX A and B FEX, are active active connections?
  2. For exampleIs blades 1 and 5 has made the fallen on the same ports on the same FEX FEX?
  3. What is the method that load balances blade ping?
  4. Assume 1 Uplink fails between FEX and FI, is the uplink 1 on FEX B overexploited.
  5. How the virtual machine decides which ANC port to send traffic?

  She now...

  See online

  When the blades connect to FEX A and B FEX, are active active connections?

  I assume you mean of FEX = IOM in the chassis;

  Yes the two fabrics are active; all VIC are double hosted the fabric a and b

  For exampleIs blades 1 and 5 has made the fallen on the same ports on the same FEX FEX?

  You have the choice between static pinning (made by UCS Manager) blade slot to an Uplink of IOM, or set a port between IOM and fabric channel interconnection; port channel is the preferred mode.

  Static pinning depends on model of IOM, 2204 / 2208 and the number of uplinks, for example

  2204-4 uplinks

  slot 1 and 5--> uplink 1 chassis

  2 and 6-> uplink 2 chassis slot

  ......

  2208-8 uplinks

  Chassis slot 1-> uplink 1

  Chassis slot 2-> uplink 2

  ......

  Pinning static if a rising fails, the fabric slots correspondent has or b, are declining. There is no repinning for remaining uplinks; grouping and multipathing will take the appropriate action.

  If you have a port-channel, and you lose a link, no problem, as long as there is at least a link to the top; your bandwidth will decrease, but th vnic/vhba are still in place

  What is the method that load balances blade ping?

  Don't understand; To be more specific

  Assume 1 Uplink fails between FEX and FI, is the uplink 1 on FEX B overexploited.

  Yes, in principle. but you will have to design oversubscription from end to end in any case!

  How the virtual machine decides which ANC port to send traffic?

  It depends on the hypervisor; Take, for example, VMware ESXi and vswitch: load balancing can be configured by switch and / or port-group; important caveat: do not hash IP as interconnection of fabric does not support the Protocol LACP; by default, the original virtual¨port code.

• Use of external ports on the fabric of the UCS of interconnections

  Hello everyone

  We have a new UCS with 2 fabric interconnects and I want to uplink the ISCSI NICS to an external server not cisco.   What I read, it looks like using a port of the device is the way to go.  However I have noticed on the port configuration of the server, he does not see the VLAN already defined.  The server communicates only with a SAN upstream on our main switch;  If you define a VLAN separately on external does this mean if my server had to communicate with a blade of Cisco low flow to send traffic to the base and then back down even though he was on the same VLAN?  Just try to understand it.  Would it not be better to configure the mode switch?   I know that you will lose some functionality with mode switch.

  Thank you

  Jason,

  I don't ' know if you've already read, but here's the doc of a connectivity device port.

  http://www.Cisco.com/c/en/us/support/docs/servers-unified-computing/UCS-...

  Thank you

  Saurabh

• N1K or DVS info on an ESXi host

  We try to find a way to see if an ESXi host is on a N1K or standard pass that we move all systems to N1K and must check it was completed.  We thought using the CDP - Get but since one uses Cisco UCS blades that keeps Cisco Nexus.  Does anyone have any ideas on how this can be done?

  Thanks in advance.

  Matt

  I don't have a handy N1K right now, but it lists the ESXi hosts connected by N1K?

  foreach($dvs in Get-VDSwitch){  Get-View -Id ($dvs.ExtensionData.Config.Host | %{    $_.Config.Host}) |   Select @{N="dvSw";E={$dvs.Name}},@{N="ESX";E={$_.Name}}
  
  }
  

• Map of ESX/ESXi DRS/HA/FT-UCS M71KR - Nexus 1000

  Hi all

  I have a doubt about the configuration of ESX/ESXi in congiunction with ucs years n1k.

  If I use the channel (mac-pinning based due to material ucs) on n1k and I have only two NIC (M71KR) card can I properly configure FT/HA or DRS.

  having a port-channel vmware is informed of nic ha? I may have some redundant warning as a note to management interface or stuff like that?

  any idea of?

  TNX

  Dan

  Dan,

  We work with the DRS, FT and HA. You will be able to configure a Port-Channel with MAC-pinning and always use DRS, the FT and HA feature.

  To create the channel port, you must add the two network interface cards to the DVS in a perspective of VMware is aware that you have redundancy NIC.

  Louis