LACP Configuration for Distributed Switch - vSphere 5.5

I'm trying to set up the switch distributed in a new lab, but I have a question on how to properly configure the port offsets.

If I have a collection of servers, each with two uplinks to a physical switch, how to properly configure the LACP lag? All ports on all of the servers are part of the same GAL, or should I create a different OFFSET for each new host?

For example:

Lets assume 16 servers. I create a single GAL on the DSwitch, mapping of all physical links 32 rising at the same OFFSET, then replicate the physical switch configuration? Or should I create 16 different gal, one for each pair of uplinks on each host server?

It seems strange to me that I should set up all servers on the same SHIFT, containing the uplinks for different physical host computers.

Aaron, looks like you got it. You can think of the LAG configuration on the switch distributed as a profile of the EEG applied to each host. The physical switch upstream is configured with an OFFSET for each host.

Tags: VMware

Similar Questions

  • Automatic configuration for routers, switches Catalyst and ASA backups

    I am looking for a free solution to make monthly backups of my routers (2821), Catalyst (X 3650, 3750-X) switches and ASA (5510). I'm in a Windows environment and have you not mind doing a bit of coding.

    I did some research looking at other popular solutions:

    -SNMP and a combination of Bash scripts, but that does not support switches Catalyst from what I've read.

    -Rance, on Linux & OS X, not something common in our environment

    -Tools of Tao kiwi, not free

    Is there something (or if applicable, somethings) that I am missing that will do this from a Windows environment for free?

    Thanks in advance.

    Kron seems to be supported on the routers only, ASA here is a good explanation on how to collect the backups regularly:

    https://supportforums.Cisco.com/docs/doc-14958

    If you are looking for a centralized solution and you machine to act as a collector, rancid is really the best option (if you can allow non-windows machines).

    Kind regards
    Ivan

  • Distributed switch

    Assuming that the new host that is added to the existing cluster is the same server, I have a question:

    If the cluster is using distributed switch, how to set up the new host to use the existing configuration of distributed switch?

    Once the host and the network/uplinks are correctly configured, simply follow steps shon in Add hosts in a Distributed Switch vSphere

    André

  • Distributed Switch vs switch Standard

    Hey all,.

    I was wondering if someone could answer some questions for me on distributed switches. I've worked with vmware for a few years now, but believe it or not I've never configured a distributed switch. The environment I work in always used standard switch, and when you're dealing with people who have the mentality of "if it ain't broke...". ', it is easy to become complacent. Everything in my environment is 5.1 incidentally.

    I read various articles on capacity now, but I don't have a test environment to break, so I have to pay attention to any changes I make. I thought it was better to ask those who are familiar with him before deciding on future changes.

    -What I lose any host on network management (what I used with switches std) as soon as I create a distributed switch?

    -Distributed is vCenter wide switching, or can I create several switches distributed by cluster, or only with certain subnets included?

    -What happens when I add a new host in a cluster with a distributed switch? It is automatically get set up with the distributed Switch networking capabilities?

    -Are there pitfalls or problems I should be aware, either when creating a distributed switch or in the day-to-day management of the/them?

    -What is the best thing that I win? Worse, I've lost?

    Thanks for your comments.

    Hello

    To answer your question:

    (1) so you don't lose everything as soon as you create the switch. You can create a vDS (assuming that licenses) at any time and configure / play with him. Until connect you hosts and uplinks, so it is a pretty useless component in vCenter.

    (2) a vDS is wide vcenter. However, you can create them individually if oyu want. That is to say - one for the traffic of comments, one for management, one for storage.  You can also have them by cluster if you REALLY want to and just add the hosts in this cluster in particular switch that you built. But reducing to zero the point of them a bit.

    (3) when you add the host, you will be asked what you want to use, if you want to migrate all the standard switch VMK interfaces and if you want to migrate VMs. ACE as soon as this host and its specific uplinks are on the switch then the host will be able to use the vDS for everything you have set for rising connections.

    (4) didn't really create them. Some people are a little cautious host management of investment (usually vmk0) on them, some people are not. In general, the rule is that with a vDS you can not manage it if vCenter is down. But with a vSwitch you can connect to the host directly and set it up in this way individually. There are traps in this respect of binding of static ports on the port groups, but it can be bypassed by the binding of ports muralism on the management ports group. Chris Wahl explains better than I've rambled

    (5) the best thing you can is to have a centralized management and create groups of ports. You have a new VLAN? Need a new group of port? Simply create once for the vDS and it applies to all guests! In addition, you also get control network IO, where you can assign actions to specific types of traffic and other goodies that are not available on standard switches. I would say that you lose much, vDS if you have the license and and a fairly large area are the way forward management. Trying to think really inconvenience but I can't!

    Happy to help you further!

  • migration of the distributed switch standard, how much should be created?

    I'm migrating my VMs from standard to distributed switches.  Are there practices for distributed switches how you create?  Vmotion traffic must be on a separate dvswitch?  My plan was to a dvswitch for vMotion and the other for my VMs and mgmt of ESXi servers.  OR would they all simply in the same switch?

    Thanks in advance.

    Do not forget that each vDS need your own uplinks (physical network interfaces)... If you have enough rising, you can create several vDS, but with a small number of uplink, a better approach is to create just a single vDS and a group of ports for each type of traffic (vMotion, mgmt, VM)... and you can specify the uplink for each port group to use by using the configuration of aggregation and failover.

  • New to distributed switch necessary Clarification.

    Hi all

    I need some clarification.

    Have created a distributed switch, in my lab, which was created without any problem.

    I have 2 hosts in my lab host A and host B.

    There is no vmkernel ports to migrate, so away I went to Setup - network - Distributed Switch - manage e-cards and created a host A vmkernel port.

    But when I goto host B Configuration - Networking - Distributed Switch, I am not able to see the vmkernel port I created.

    I need to create the port on host B manually vmkernel group? I was under the impression, the port of core, I created on host A would be automatically visible to host B?

    That's where I need some clarification, so that, from now on, I am being harassed by this.

    Any clarification would be appreciated.

    Thanks Andrew for the clarification, much appreciated and its clear now.

    Port Comon group

    Core groups (e-cards) must be configured individually on each server.

    Do not crack.

    Thank you once more for confirmation

  • 'upgrade' control of Distributed Switch for network IO / LACP improvements via PowerCLI?

    vSphere 6.0.  PowerCLI 6.0 R1.


    After using the New-VDSwitch PowerCLI command, I noticed that the Web Client vSphere shows there are "updates available" to the distributed switch resulting.


    I know that I can right-click on the distributed switch and select upgrade in the Web Client.


    Someone knows how can I perform the upgrade using PowerCLI?

    Capture.PNG

    search terms: nioc

    Here is the code I ended up with yesterday before reading your response... is it ok for you?

    Also, is it OK to use 'ReconfigureDvs' instead of 'ReconfigureDvs_Task '?

    PS how do you get syntax colorization of your code PowerCLI?

    $myvDS = news-VDSwitch-Verbose - name $myvDSwitchName - location $myNetFolder - NumUplinkPorts 2

    # Upgrading capacity DSwitch "NIOC v3" and "Enhanced LACP Support.

    $spec = new-Object VMware.Vim.VMwareDVSConfigSpec

    $spec.networkResourceControlVersion = 'version 3 '.

    $spec.lacpApiVersion = "multipleLag".

    $spec.configVersion = $myvDS.ExtensionData.config.configVersion

    $myvDS.ExtensionData.ReconfigureDvs ($spec)

    # Activate the NIOC

    $myvDS.ExtensionData.EnableNetworkResourceManagement ($true)

    # Extract our object to collect the current configuration

    $myvDS = get-VDSwitch $myvDS

  • VSS migration to virtual Distributed Switch configuration

    Hi all

    I am trying to wrap my brain around that and just run into a few problems actually make things work. Please bear with me, I will try and describe the environment that I have and what I'm trying to building with like jargin little I can.

    My current vmware environment consists of 3 hosts vsphere 4.x and about 6 different subnets. My primary host vSphere is home to the largest part of the virtual machine and manages 5 different subnets connected to each of 5 virtual switches separated with 1 assigned to each NETWORK adapter. Also, there is a switch of kernel VM with a connection to my NetApp iSCSI. The other vSphere hosts are simple enough, the two are connected to subnet 1 with a virtual switch for it and a switch of kernel VM with a connection to the NetApp iSCSI.

    Each subnet in my lab is managed/break through the VLANS on Cisco devices, so I saw that it had to assign any settings VLAN since the power of VMware.

    If you refer to the VMWare vNetwork Distributed Switch: Migration and vmware Configuration guide, I am trying to migrate a seup similar to this:

    vmware_multiple_vds.JPG

    However, I'm running issues when you try to get the int hosts a vDS configuration. I could create a vDS for my root subnet, add one of my hosts vSPhere and migrate the virtual machine to the new port group in this vDS. The Service console as well as the VMKernel remain virtual switches on the host and I can't understand how these migrate to a vDS host without lose the connection.

    I'm asssuming based on the number of subnets that I manage between hosts, I'll finish with about a 5-switch vritualDistributed for subnets, each with at least 1 card a vSphere host physical NETWORK link up to the appropriate subnet. In addition to this, I'm assumining I'll need to create a vDS for the Service console and VMKernel (iSCIS) traffic. The Service Console are on the same subnet, some VM most residence on that subnet, separate VSS was created on the hosts to manage separate traffic.

    Any help anyone can provide on how to create vDS for SC/vmk traffic and get the associated host migrated to which would be very useful. As I said I was able to create 1 vDS and add a host computer via 1 uplink NIC with the virtual machine, but nothing beyond that seems bad connection.

    Feel you please free to ask for additional details, I know it's a lot and maybe a bit confusing. Thank you.

    -Bryan

    Hello

    If I'm correct, when the host is added the vDS and the creation of the vDS Service Console, I have to select an unused NIC and migrate the SC existing to the new group of port... or should I add a second SC for the VSS by using a NETWORK card available and who migrate to the new vDS?

    Yes, select an unused physical nic so that you have a physical nic connected to the VSS and the other to connect to the uplink of vDS group and migrate the SC. existing if you have several hosts, vMotion all VM and test to see if the migration without distruption, also works to keep details of the ILO ready incase you need to connect to the console of the server.

    This is a very good book white http://www.vmware.com/files/pdf/vsphere-vnetwork-ds-migration-configuration-wp.pdf that provides detailed information about the migration.

    All the best.

    Kind regards
    Arun

    If you have found this or other useful information, please consider awarding points to 'Correct' or 'useful '. Regards, Arun VCP3/4, HPCP, HP UX CSA http://kb.vmware.com/

  • How to configure vSwitch distributed for EQL MPIO?

    I read the guide of Dell for the vswitch for EQL configuration. However, it is assumed that DVswitch is not used. Does anyone have guide how do for distributed vswitch?

    What version of ESX?   MEM 1.1.x can create the configuration and it supports the VDS.

    Refer to TR1074 MEM 1.1. install the report tech for more information.

    In addition, TR1075 the configuration of vSphere guide has instructions on Configuration vDS.  So I'm not sure which document you were using.

    What 10GbE and 1GbE?    If you have distinct physical maps, I would still use a standard vSwitch on vDS.   Not really convinced that there is an advantage.

  • vSphere Distributed Switch

    Hi, I recently updated my vcenter of 5.1 to 5.5 but my vSphere Distributed Switch (VDS) is always version 5.1. Now, I have to create another VDS is it compatible with the vcenter 5.5 or do I need to update my

    VDS from 5.1 to 5.5

    Hello

    Please check the link for vDS below set to level.

    https://pubs.VMware.com/vSphere-55/index.jsp?topic=%2Fcom.VMware.vSphere.networking.doc%2FGUID-330A0689-574A-4589-9462-14CA03F3F2F4.html

    In brief

    You must have first the vCenter to 5.5

    Then, ESXi 5.5 using the vDS

    Then you can upgrade vDS on the fly.

    Suhas

  • Configure VTEP VXLAN on distributed switch?

    Hello everyone

    I've found this discussion: How to test VXLAN

    But I'm still not sure.

    I environment vSphere 6 with a 3-node cluster and server vCenter (trial mode).

    I would just try the VXLAN by adding a VTEP to a distributed switch.

    For each of the nodes would be a VTEP.

    I really should he use Manager of NSX and NSX controller etc. to achieve this, or is there the possibility to simply add a VTEP to a distributed switch?

    KR & thanks in advance!

    Lukas

    You can use the last Director 5.5.3 vCloud and vCloud network and security 5.5.4, they are compatible and supported for ESXi/vCenter 6.0.

    See VMware product interoperability Matrices or documentation of the respective product for more details.

  • vSphere Distributed Switch 5.5 traffic filtering and tagging

    Someone had a chance to create a script to update the traffic filtering and marking of area of a group of ports on a vSphere 5.5 Distributed Switch? The settings are only exposed in the web client for Onyx is not an option.

    I need to create a rule with the values below.

    Traffic Filterig and marking:

    Set State enabled

    New rule of network traffic

    Name: name of the traffic rule

    Action: Tag

    CoS value: tag value Update CoS: 4

    Qualifiers of traffic:

    Traffic management: evacuation

    New qualifying traffic system: vMotion

    This is as much as I can get.

    $VDSPortGroup = get-VDSwitch Test - dvSwitch | Get-VDPortGroup Test-PG

    $Spec = new-Object VMware.Vim.DVPortgroupConfigSpec

    $Spec.configVersion = $VDSPortGroup.ExtensionData.Config.ConfigVersion

    $Spec.defaultPortConfig = new-Object VMware.Vim.VMwareDVSPortSetting

    $Spec.defaultPortConfig.FilterPolicy = new-Object VMware.Vim.DvsFilterPolicy

    Sorry, it took a little longer than expected.

    Try like this

    $dvSwName = "dvSw1".

    $dvPgNames = "dvPg1".

    $dvSw = get-VDSwitch-name $dvSwName

    # Activate LBT

    foreach ($pg in (Get-View-Id $dvSw.ExtensionData.Portgroup |)) Where {$dvPgNames - contains $_.} {Name}))

    $spec = new-Object VMware.Vim.DVPortgroupConfigSpec

    $spec. ConfigVersion = $pg. Config.ConfigVersion

    $spec. DefaultPortConfig = New-Object VMware.Vim.VMwareDVSPortSetting

    $spec. DefaultPortConfig.FilterPolicy = New-Object VMware.Vim.DvsFilterPolicy

    $filter = new-Object VMware.Vim.DvsTrafficFilterConfig

    $filter. Nom_agent = "dvfilter-credits-vmware.

    $ruleSet = new-Object VMware.Vim.DvsTrafficRuleset

    $ruleSet.Enabled = $true

    $rule = new-Object VMware.Vim.DvsTrafficRule

    $rule. Description = "name of traffic rule".

    $rule. Direction = "outgoingPackets."

    $action = new-Object VMware.Vim.DvsUpdateTagNetworkRuleAction

    $action. QosTag = 4

    $rule. Action += $action

    $ruleSet.Rules += $rule

    $filter. TrafficRuleSet += $ruleSet

    $spec. DefaultPortConfig.FilterPolicy.FilterConfig += $filter

    $pg. ReconfigureDVPortgroup ($spec)

    }

  • Configuration of VLAN Switch Distributed

    Hello

    This a configuration problem and I'm not really sure how to set it up.

    I created a distributed switch, and ESXi1 and ESXi2 are members. I created a comeback portgroup named A_01 and is a member of the VLAN 101.

    I created a virtual XP1 machine in ESXi1 located in A_01 and I created a VM XP2 on ESXi2 located in A_01.

    They do not communicate.

    I have a switch between them and the uplink for the distributed switch of ESXi1 is connected to port1 and ESXi2 uplink is connected to port2

    I have 'tag' port1 and port2 in vlan 101.

    They do not communicate.

    I have both change the network of the VM XP1 and XP2 VM a portgroup with no VLAN they communicate.

    What I am doing wrong.

    Thank you!

    "I have both change the XP1 XP2 VM VM network in a portgroup with no VLAN they communicate."

    --> its because your physical switch is configured as port access . Marking is done at the level of the physical switch. This is the expected behavior.

    If you want to tag the VLAN level vSwitch, then you must put the physical switch as a trunk port and VLAN 101.

  • Error HTTP 500 HOL - vSphere Distributed Switch from A to Z-

    It seems that there are problems with the "vSphere Distributed Switch from A to Z" HOL.  I get the following error message when you try to access the laboratory (see screen capture below).  Other labs are working fine for me.  I tried to access this laboratory of 5 different computers with the same results.

    There was a communication error with the server. Try again in: 26

    Message: HTTP 500 Code

    Lab Console - VMware NEE - Google Chrome_2013-09-30_10-41-50.png

    Any help would be appreciated.

    Try the testdrive tenant.

    He seemed to do a pod stuck in the queue of our deployment for the testdrive tenant. That may have caused this problem.

    -Doug

  • Consolidation and failover for the uplink on the Distributed switch port group

    Hello

    I have a problem with the implementation of a distributed switch, and I don't know I'm missing something!

    I have a few guests with 4 of each physical cards. On the host eash I configured 2 virtual switches (say A and B), with 2 physical network by vSwitch using etherchannel adapter. Everything works fine for etherchannel and route based on the hash of the IP for the latter.

    Recently, I decided to create two distributed switches and move the respective physical ports of virtual switches to this distributed switches. Once again, I want to configure etherchannel and route based on the hash of the IP. But when I open the settings for the uplink port group, aggregation and failover policies are grayed out and cannot be changed. Apparently they inherit configuration also but I don't know where!

    Chantal says:

    Once again, I want to configure etherchannel and route based on the hash of the IP. But when I open the settings for the uplink port group, aggregation and failover policies are grayed out and cannot be changed. Apparently they inherit configuration also but I don't know where!

    You must set the card NIC teaming policy on trade in reality and not on the uplink group more expected.

Maybe you are looking for