UCS and QoS

Hello:

The ecosystem of the UCS relies on a solution of aggregation of port to the i/o chassis, namely, FEX modules.

The FEX modules are not fully featured switches. Nor do they have a political intelligence transfer at all. Instead, the FEX modules deploy approach "pinned" in which the descenders (those facing NIC in the Server Blade, LOM, mezzanine cards) are mapped to an uplink port (those who face a 6100 Interconnect fabric) to form what can be described as a group of aggregator.

The result is a simplified approach to blade i/o in which traffic patterns are predictable and failover is deterministic. In addition, there is no need to configure STP because ports are sent in order to exclude any possibility of a loop connection.

This having been said, there is some merit to the argument that this port aggregation design places a hole in the middle of a deployment of QoS for the scheduling of packets on the uplink to the 6100 of interconnection fabric ports is not done in a way that recognizes the priority?

In order to develop a little more, one can have a VMware deployment and use NetIOC or maybe configure QoS on a 1000v switch (which uplink ports are mapped to a port on the VIC de Palo) and configure QoS on the VIC and then the 6100 Interconnect fabric. But, given that the FEX is not planning for traffic to the 6100 interconnection fabric on a priority basis, the deployment of QoS has a hole in the Middle, so to speak.

Thoughts?

Hello

I posted an answer to your question here:

http://bradhedlund.com/2010/12/08/Cisco-UCS-fabric-Extender-FEX-QoS/

I hope this helps.

See you soon,.

Brad

Tags: Cisco DataCenter

Similar Questions

  • Upgrade of CUCM 7.1.5 to 9.1 (contract No. UCS and ESW)

    Hello

    We have CUCM 7.1.5 cluster (two servers) without contract UCS and ESW.

    What would be the better and way cheaper to upgrade to CUCM 9.1.

    We intend to use our VMWare system and only buy the CUCM 9.1 software (not servers).

    The current DLU is trensfered or should we buy everything from scratch.

    Kind regards.

    Demont - my friend, how have you been? I hope you do well!  Very nice to see you

    always help others - keep it! (+ 5)

    Kasper,

    As my friend that Demont said accurately, you contact your partner Cisco Channel/Account Manager. I just wanted to add, you can migrate DLU, but like with CUCM9, you have only UCL or CUWL - according to the type of license you select UCL/CUWL, you will need to purchase additional licenses and buy the UCS. But having said that you can benefit from Cisco in the car at 9 and can buy this at a really discounted price. Your Cisco representative can explain all the options in detail.

    Thank you

    Terry

  • UCS and L2 multicast issues

    Hello

    I have 2 data centers connected through a tissue of spine / sheet using BGP EVPN. In a domain controller, I UCS and other legacy Dell ESX servers connected to catalyst switches. These are shared resources via a VPC local 9 k leafs to interconnect with the fabric. Some of VLANS have L2 multicast running. All the VLANS relevant are bridged through the fabric and presented to UCS/ESXi. When virtual machines have been vmotioned overall everything has worked well except the multicast broke. I created a policy of multicast in UCS and activated the interrogator with an IP address in one of VLAN relevant from the cisco documentation what UCS will not pass traffic unless there is a mrouter/interrogator. Then, I joined politics to the Vlan relevant under the section of LAN. However still does not. I don't see any questioning on the directly connected switches information upstream is. Should applicant IGMP allowed on switches?

    Thank you!

    https://supportforums.Cisco.com/document/12725606/setting-multicast-with...

    https://supportforums.Cisco.com/discussion/12111706/UCS-and-multicast

  • UCS and storage network

    I am using iSCSI with the software iSCSI adapter in ESXi as well as of the UCS and use binding to ports to bind each vmkernel iSCSI port to a specific vmnic and I use 2 vmnic.  Must I also within the UCS links each vmnic so that vmnic1 traffic to cross a 1 interconnection fabric and the movement of vmnic2 fabric crossbar Interconnect 2?

    Must I also within the UCS links each vmnic so that vmnic1 traffic to cross a 1 interconnection fabric and the movement of vmnic2 fabric crossbar Interconnect 2?

    A vEth can only come out in a unique fabric of interconnection - it's decided when you take the fabric (A or B) for the vNIC. There is no additional connection to perform.

  • VM ESXi and Cisco's UCS and EMC without a question of a novice

    I'm confused about the places where to install ESXi 5.0. This is on a windows Server or UCS blades if you have blades of the UCS and EMC San

    I know that we must install vcenter, a first round upward on the Cisco UCS blades. Do I connect EMC San?

    How to set these at all? I need a tutorial step by step for these 3 types of equipment to set up.

    Thank you!

    You can install any OS on blades, it's simple as console for all blades connected with the network and storage, it requires knowledge networking and storage configuration Cisco blades. Each blade has a remote as ILO or DRAC console, you can mount ISO image and start to install the OS, according to the needs we can configure boot from the local disk or boot from SAN, engineer of the UCS who has enough knowledge will create profiles of service for a whole different requirement for Windows Server and Citrix Xen server or ESXi servers , each requirement will be have the profile, just add blades and card service profile to access the console Blade Server, please download UCS virtual device and confiure and start learning.

  • Limit my bandwidth downloading the applications using the API to control traffic and QoS

    I used QoS and Traffic Control API as TcAddFlow and TcAddFilter to control my bandwidth usage download applications.

    We manipulate TC_GEN_FLOW, to send and receive FLOWSPEC parameters.

    Now, I want to set the exact limit to 5 Mbps, what are the value that I need to set for TokenBucketSize and TokenRate to limit bandwidth to 5 Mbps FLOWSPEC structure?

    Code snippet:

    newFlow-> ReceivingFlowspec.DelayVariation = QOS_NOT_SPECIFIED;
    newFlow-> ReceivingFlowspec.Latency = QOS_NOT_SPECIFIED;
    newFlow-> ReceivingFlowspec.MaxSduSize = QOS_NOT_SPECIFIED;
    newFlow-> ReceivingFlowspec.MinimumPolicedSize = QOS_NOT_SPECIFIED;
    newFlow-> ReceivingFlowspec.PeakBandwidth = POSITIVE_INFINITY_RATE;
    newFlow-> ReceivingFlowspec.ServiceType = SERVICETYPE_NETWORK_CONTROL;
    newFlow-> ReceivingFlowspec.TokenBucketSize = ?;
    newFlow-> ReceivingFlowspec.TokenRate =?;

    newFlow-> SendingFlowspec.DelayVariation = QOS_NOT_SPECIFIED;
    newFlow-> SendingFlowspec.Latency = QOS_NOT_SPECIFIED;
    newFlow-> SendingFlowspec.MaxSduSize = QOS_NOT_SPECIFIED;
    newFlow-> SendingFlowspec.MinimumPolicedSize = QOS_NOT_SPECIFIED;
    newFlow-> SendingFlowspec.PeakBandwidth = POSITIVE_INFINITY_RATE;
    newFlow-> SendingFlowspec.ServiceType = SERVICETYPE_NETWORK_CONTROL;
    newFlow-> SendingFlowspec.TokenBucketSize =?;
    newFlow-> SendingFlowspec.TokenRate =?;

    Thank you & best regards

    This issue is beyond the scope of this site (for consumers) and to be sure, you get the best (and fastest) reply, we have to ask either on Technet (for IT Pro) or MSDN (for developers)

    If you give us a link to the new thread we can point to some resources it
  • FlexVPN and QoS on tunnels

    There is a simple topology: a hub and spoke. FlexVPN is woking together with psk, BGP, and no RADIUS.

    Now I want QoS on the hub and the spokes. The Center has an ISP connection, let's say 100 MB and some rays have 10 MB, some 5 MB and so on.

    Each ray has a tunnel interface and a virtual-template interface. I can apply "service-policy output" on these interfaces, no problem. (Should I apply "service-policy output" on the tunnel or on the interface virtual-template interface or both of them, I'm still not sure, but this isn't a big problem)

    What should I do with the hub that does that one tunnel interface and a virtual for all model the rays?

    If I had 100 spokes hub would still have only a single tunnel interface and a virtual model for all the rays. The hub also has virtual-access interfaces for each Department, they sort of dynamics, I do not create them, they appear by themselves and I am not able to configure. When I try to configure the Cisco says: % Please use virtual model to configure your virtual access.

    Where and how I can apply 'service-policy output' on the hub so I want unique QoS for each Department?

    Given that you use no RADIUS, you can apply config dynamically with AAA attribute lists.

    I described the similar config (including a very basic policy) in this document http://www.cisco.com/c/en/us/support/docs/security/flexvpn/116032-flexvp...

    To answer your questions, you always apply the config to go through the model.

    (In this case) Attributes are added to the don't go to STM, you use VT as a basis for what you need, followed by additional dynamic attributes for SAV.

    For interfaces tunnel (on the shelves), it is quite easy to enable QoS, but what you could look on the wall policy on the physical interface and not the tunnel interface (do not forget that the DSCP values are copied on to external header). After all, you want to manage the bandwidth to ISP not to cloud VPN, most of the time.

  • SKUs recovery Cisco UCS and expired ESW

    I have a client who has expired ESW, but still the UCS. I got the word that there is a recovery sku (L-REINST-UWL-STD), but is not in the CCW. Anyone can shed some light on this?

    I suggest that you take a look at the v9 OG CUWL, you will need to use the highest level SKU:

    L CUWL-MISC

    I just tested it and it works

    Reference:

    http://www.Cisco.com/Web/partners/downloads/partner/WWChannels/technology/IPC/downloads/finalcopy.PDF

    Best regards

    Jaime

  • bandwidth and QOS

    Hi guys,.

    I get a line of lease 20Mbps between two offices and it connects two cisco C4507R switches. I have configured the QOS on the two switches, and I know the QOS will take effect when the network congestion occurs. But the ports that connect the rental displays 100 Mbps on the switch. So I configured 'bandwidth 20480' command in the ports, if this will help active QOS when the network stream is up to 20Mbps?

    my command under the interface:

    interface GigabitEthernet1/38

    No switchport
    bandwidth 20480
    IP 10.81.16.4 255.255.255.248
    service-policy output QOS - SH

    Disclaimer

    The author of this announcement offers the information in this publication without compensation and with the understanding of the reader that there is no implicit or explicit adequacy or adaptation to any purpose. Information provided is for information purposes only and should not be interpreted as making the professional advice of any kind. Use information from this announcement is only at risk of the reader.

    RESPONSIBILITY

    Any author will be responsible for any wha2tsoever of damage and interest (including, without limitation, damages for loss of use, data or profits) arising out of the use or inability to use the information in the view even if author has been advised of the possibility of such damages.

    Poster

    Your 4500 QoS will only engage when the interface clutter.

    What you need is a shaper with QoS support, that can match bandwidth your provider.

    Unfortunately, this is not a feature of the 4500 series.

  • Nexus 1000v, UCS, and Microsoft NETWORK load balancing

    Hi all

    I have a client that implements a new Exchange 2010 environment. They have an obligation to configure load balancing for Client Access servers. The environment consists of VMware vShpere running on top of Cisco UCS blades with the Nexus 1000v dvSwitch.

    Everything I've read so far indicates that I must do the following:

    1 configure MS in Multicast mode load balancing (by selecting the IGMP protocol option).

    2. create a static ARP entry for the address of virtual cluster on the router for the subnet of the server.

    3. (maybe) configure a static MAC table entry on the router for the subnet of the server.

    3. (maybe) to disable the IGMP snooping on the VLAN appropriate in the Nexus 1000v.

    My questions are:

    1. any person running successfully a similar configuration?

    2 are there missing steps in the list above, or I shouldn't do?

    3. If I am disabling the snooping IGMP on the Nexus 1000v should I also disable it on the fabric of UCS interconnections and router?

    Thanks a lot for your time,.

    Aaron

    Aaron,

    The steps above you are correct, you need steps 1-4 to operate correctly.  Normally people will create a VLAN separate to their interfaces NLB/subnet, to prevent floods mcast uncessisary frameworks within the network.

    To answer your questions

    (1) I saw multiple clients run this configuration

    (2) the steps you are correct

    (3) you can't toggle the on UCS IGMP snooping.  It is enabled by default and not a configurable option.  There is no need to change anything within the UCS regarding MS NLB with the above procedure.  FYI - the ability to disable/enable the snooping IGMP on UCS is scheduled for a next version 2.1.


    This is the correct method untill the time we have the option of configuring static multicast mac entries on
    the Nexus 1000v.  If this is a feature you'd like, please open a TAC case and request for bug CSCtb93725 to be linked to your SR.

    This will give more "push" to our develpment team to prioritize this request.

    Hopefully some other customers can share their experience.

    Regards,

    Robert

  • Masking vs. zoning for UCS and NetApp storage LUNS

    I know it's more a matter of Cisco MDS/storage, but nobody knows masking the LUNS and zoning, which one would be the most preferred method? I have two switches Cisco fabric 9148, two controllers SAN Netapp FAS3210 1 UCS chassis with 4 blades B200 M2. I said that I should not connect the Cisco UCS fabric interconnect directly at the back of the NetApp SAN and configure with LUN masking, but I rather configure a fabric zoning on switches Cisco 9148 MDS. Normally, this wouldn't be a problem but we have an offsite location where we have not all DMS switches to, and I would like to connect them directly to the San.

    I was told that this could lead to corruption if misconfigured dsik and point of view of Cisco is to use a zoning by some type of switch Cisco Fabric. Of course, I have of course this Cisco advises anyone to manage this type of configuration through their facilities instead of on the SAN. Does anyone have an opinion on the matter?

    Zoning and masking are two completely different characteristics.

    Zoning occurs on your storage switches and is the equivalent of an ACL (Access Control List).  It limits who may be considering other targets and/or initiators.  (Who I see?)

    Occurs on your storage array and masking limits what initiator LUN has access to.  (What do I see?)

    * UCS is not supported to directly connect a storage array in the interconnection of fabric, at least to have a storage upstream switch (MDS or equivalent) to push the zoning.  The use of zoning prevents a faulty initiator potentially impacting on the operation of the others by limiting what they can see in the fabric.

    If you happen to have a Nexus 5 K by chance, they can also work as your storage switch.  The N5K is able to perform almost all of the same services as MDS fabric and is fully supported.

    Kind regards

    Robert

  • The review of the UCS and VCP

    Hi all

    I do not have a secure Channel and yet so I will at least make the test of the UCS?

    Is-Cisco has a system as they will allow us to test of the UCS, but they will not give us unless we are SECURE.

    Thank you

    Ismail

    Hi Ismail

    You can always take the test and still qualify as a Cisco Unified Computing Technology Specialist of the design and a Cisco Unified Computing Technology Support Specialist without a PCV.

    But if you're after Cisco Data Center Unified Computing Support Specialist, etc. then yes VCP is always a pre - req (I heard that there will be rejected in favour of a managed Cisco equivalent, but who is without merit)

    Unfortunatley (that I found) just passing examination VCP wasn't enough, but the VMware course is also a mandatory requirement for VCP VMware.

    Good luck with it.

    Concerning

    Colin

    http://ucsguru.com

  • Implementation of VLAN and QoS for VOIP on SG200-18

    We recently purchased the smart switch SG200-18 to replace a Netgear switch. We are moving our phone service to VOIP through our local ISP as well.

    I currently have the VOIP phone plugged into Port 17 on SG200-18 (it is a Grandstream Cordless VOIP phone).

    I want to put the VOIP phone on one VLAN separate from the rest of the network and optimize QoS parameters so that the VOIP phone has exceptional audio quality even during network traffic.

    Here are my questions:

    1. do I need to set anything on the type of port to Port 17 (because it resembles a shape any Combo port)?

    2. How can I do to isolate VOIP telephone it's own VLAN (I see the parameters VLANS and VLAN voice, not sure that one to use;) I've tried to set a VLAN and broke the Internet connectivity on the phone until I went and removed)?

    3. do I need to adjust the QoS settings to switch to better optimize the VOIP phone?

    Some additional questions about the GS200-18 in general:

    1. do I need to adjust the parameters of the system on the switch time? I am in the Central time.

    2. do I need to adjust the Green Ethernet/Energy Saving parameters or should I stay with the default settings?

    In addition, a couple of "getting started" questions for Cisco:

    1. I registered an account My Cisco. What should I do to register my switch with Cisco and associate with my My Cisco account?

    2. What are the benefits of purchasing a contract of Cisco Small Business support, and how much would it cost the SG200-18 (I ordered it from Provantage)? I'm curious to see if it's worth the money.

    Here's my 'features ':

    Switch: SG200-18

    VOIP phone: Grandstream DP715 and 710 handsets

    Plugged in: Port 17 on SG200-18

    Services: Internet Local (Direclynx)

    Type of connection: 3 m down / 500 k up DSL move to a future wireless connection that will give us higher speeds

    Backend VOIP provider: VOIP Innovations

    Router: Apple Airport Extreme AC model (all Macs and iOS devices and the OS X Server on the network, so I use the Apple router facilitates installation, because is not QoS, trying to QoS and VLAN in the switch)

    Thank you all!

    Hello

    I'll just go to the list again:

    1. sounds good in the port from the drop-down list. So can I just connect the VOIP phone and go with it, correct?

    Yes, just plug in ethernet combo port and it will work.

    2. is not an issue, but I agree, Apple likely isn't compatible QoS or VLAN.

    3. thanks for the info on time/NTP settings. If I wanted to go there and try to configure NTP, how much is it and what I have to do? I want to I can give it a quick try.

    To Setup NTP on the switch is quite simple.  Go to Administration > Time Settings > time system and check the boxes to activate the main clock Source (SNTP)

    Then go to the settings of the SNTP page and add a new entry with the IP address of an NTP server.  There is a list of available NTP servers here:

    http://www.pool.ntp.org/en/

    You must also ensure that the switches Administrative default gateway is set correctly (it must be set the to the default gateway, probably the most convenient airport) so the switch can contact the NTP server.  That option is set under Administration > Interface Management > Interface IPv4.  Change the user-defined default gateway and enter the IP address of your airport (or whatever your default gateway for your network)

    4 sounds good on the Green Ethernet settings. I'll leave it as default value.

    Yes, better to just let those unless you have weird problems with ports disconnect, who can sometimes be caused by Green Ethernet, but if there's nothing like leave it on and save a few watts.

    5 sounds good on does not need to attach my passage to my Cisco account. Should I fill out a form any registration of the product with Cisco before calling support?

    It is not a record for support.  The only thing we need you to do is to create a Cisco account, but you have already done this, so if/when you call in support, you just need your ID for Cisco (also called a CCOID sometimes) and the serial number of your switch.

    6. thanks for the info on the Service contract. Is it something that I would need to order directly from Cisco or I who would get my Cisco partner (Provantage)? After the three years is up, treat yourself to renewal or it just falls? Is there a certain amount of time I have to buy the Service Contract forward make me ineligible?

    Support contracts are purchased through a partner Cisco, or you can get them online for the CDW or Newegg for example.  Basically, you have until the expiry of your current aid for the purchase of a new contract.  For example, right now your switch comes with 1 year of technical support.  You can only buy a contract while it is still active.  Once your three-year contract is about to run out, you're in the same situation.  You can renew it before it expires, however if you leave is up, you will not be able to put a contract on it.  Contracts are not my specialty, however, so you can check with your partner for complete details.

    7. sounds good to how data use VOIP calls. His dislikes too. :-)

    I agree, a voice call is not much traffic.  What you have described you probably don't have problems, although of course I can't guarantee that.

    8. because it is from your provider and they specifically mentioned the VOIP, I would say that you'll be fine here.

    You had also placed on your airport using access point behind a router in small businesses.  I would like to say that it is possible, a large number of wireless routers have an option to put access point only mode or something like that, but you should check with Apple on how to do it.

    Insofar as a Small Business router if you decide to upgrade for the options VLAN or QoS, I would recommend the RV180, or perhaps the RV320.  Two of these models are available with or without wire depending on what you decide to do with the airport.

    I think I got all the questions, but if not just let me know,

    Christopher Ebert - Network Support Engineer

    Cisco Small Business Support Center

    * Please note the useful messages *.

  • Routers RV: personalized Services and QoS

    I have a few questions about the addition of personalized services and their use in QoS for RV routers (I use the RV220W)

    Suppose I have create a personalized as:

    Name: MyService

    Port: TCP 60000

    1 this port number is the port number used by my computer on the local network or the port number on the other side for example a computer on the Internet, I use this service to connect to? Some services use the same port number on the side that starts the service and the receiving end, but there are services that use different ports for the end of the initiator and receiver. So what is the significance of the port number when creating a custom service?

    Computer in the local network: (port TCP 60000) <---->Internet computer: (some the TCP port, not necessarily 60000)

    -or-

    Computer in the local network: (some the TCP port, not necessarily 60000) <---->Internet computer: (port TCP 60000)?

    2 suppose I have create a binding configuration of QoS profile for this service customized to a specific IP address on my local network. QoS is only applied on the outgoing network traffic. How will work this profile?

    A. applied to traffic from my LAN device with the IP address and the port TCP 60000

    -or-

    B. applied to traffic from my LAN device with the IP address specified to the port of the computer another TCP 60000?

    Thanks in advance

    Sent by Cisco Support technique iPad App

    Panos, it should be your example.

    -Tom

  • UCS and VMWare-VNIC-VMNIC active-active/active-standby

    Dear team,

    We have 3 ESXi hosts each vSwitch1 and portgroup with the name TEST with 50 of VLAN
    We have 3 Configuration of virtual machines with 19.168.0.X series and are assigned to the VLAN TEST

    Each vSwitch1 is assigned with 2 vmnic on the same VLAN 50 (from UCS vNIC, service profile - where in 1 vnic comes from the fabric and other fabric B)

    (1) inter-VM communication will work if we do an active Vmnic and other relief

    Inter 2) VM communication may work some time or may not work if we do the two active vmnic

    The above behavior is the same for virtual machines in combination with ESXihosts, expect if these virtual machines are on the same host ESxi interVm communication works very well...

    What setting to do if we do get active vmnic and always inter-VM communication

    NB:

    As mentioned above these vNIC is created from the UCS Service profile, where, in 1 vnic comes from the fabric and other fabric B, and when we try the standby active mode the interVm communication works and even when we FI's down, the standy adapter automatically rises and traffic goes well

    Thanks and greetings

    Jose

    Same condition still - the link between the FI and FI B does not serve for any traffic data. So if your source is the use of FI and your destination is on FI B, then the traffic has to go through the switch upstream. If you are still not sure, have active vmnic0 on a host computer and vmnic1 as active on the other host - communication will not take place (assuming that they go to different financial institutions)

    The solution to your problem is to define VLAN 50 of upstream network. If you don't want to do that, then you must limit all traffic within the same IF (definitely not recommended).

    Thank you

    Shankar

Maybe you are looking for