The vSphere Distributed Switch configuration on some hosts differed from that of VMware vCenter- "Out of Sync"
search cancel

The vSphere Distributed Switch configuration on some hosts differed from that of VMware vCenter- "Out of Sync"

book

Article ID: 317901

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

  •  This banner alert in the Networking view is seen:
The vSphere Distributed Switch configuration on some hosts differed from that of the vCenter Server.
  • When viewing a list of hosts attached to the distributed switch, or from the manage hosts wizard under the vDS Status column, a yellow warning icon and the message below are seen:
Out of sync



Cause

Although Distributed Switches are a component within vCenter Server, most of the switch's underlying functionality is handled at the ESX/ESXi host level. As such, relevant vDS configuration information is stored locally on the host as a proxy copy and is periodically synchronized with vCenter Server.


If network connectivity is interrupted between the vCenter Server and one or more hosts, a synchronization interval may be missed resulting in this alert being displayed. This type of interruption can occur during vCenter Service restarts, vCenter Server reboots as well as ESX/ESXi host reboots or network maintenance. This message can also occur if the host is taken out of a vCenter Server while still attached to a vDS, then brought into a new vCenter Server. Even if the vDS on the new vCenter Server is the same or similar to the vDS on the previous vCenter, the new vCenter Server will not recognize the host's proxy copy of the vDS.

Resolution

Most of the time, this message is benign and can be safely ignored until a maintenance window. Within several minutes, the host's Distributed Switch information should synchronize with vCenter Server, and the warning may clear on its own.

Process to synchronize the host vDS information from the vSphere Client:

To manually synchronize the host vDS information from the vSphere Client, follow the steps below:

  1. In the Inventory section, click Home > Networking.

  2. Select the vDS displaying the alert and then click the Hosts tab.

  3. Right-click the host displaying the Out of sync warning and then click Rectify vNetwork Distributed Switch Host.

Process to clear host's proxy copy from the host:

If the Rectify button does not clear the message or the button is not available on the host, the host's proxy copy must be cleared from the host. To do this, the host must be removed from the vDS entirely. After removal, it can then be brought back in to the vDS and the message will clear. This involves a networking change, so it is recommended to do this step during a maintenance window, if possible. 

Impact/Risks:
Ensure that MTU settings on the vSwitch and vmkernel ports match the distributed switch. Failure to have consistent MTU settings will negatively impact traffic and may cause outages for some services.

Please be aware of the following items prior to continuing:

  1. Do not migrate NSX vmkernel ports. Please follow  Command Line (CLI) migration of vmk0 from NVDS to vSS in an NSX-T environment, for the correct process to migrate between distributed and standard switches. 
  2. If LACP is configured on the physical switch, and the NICs currently in use by the host are also in LACP configuration (meaning they are in a LAG on the vDS) the LACP configuration on physical switch and virtual network will need to be temporarily broken to remove a NIC from the vDS; in order to be available for the ESXi host to use the NICs on the standard switch (unless non-LACP NICs are available and configured to pass the traffic). If the LACP configuration is not broken before moving the NIC off the vDS, this can cause further issues to the environment's connectivity.
    This will require engagement from network team managing upstream switches to change the port config. Please review LACP Support on a vSphere Distributed Switch for more details on LACP.
  3. If the physical NICs are in an EtherChannel in the physical network and the vDS portgroups are using the load balancing policy IP Hash, as long as the Standard Switch portgroups has the IP Hash load balancing policy the EtherChannel NICs can be used by a standard switch. Please review Configure Virtual Switch with an EtherChannel (or port channel) for more details on EtherChannel.

To prepare the ESXi host for the process of clearing the proxy copy, please refer to the steps below:

  1. Move all attached virtual machines to another host or to a standard switch. 

    1. A new standard switch can be built using one of the physical adapters that passes the correct VLANs from the vDS. See the note below on Load Balancing information.

  2. Move all vmkernel adapters (vmk) from the vDS to a standard switch. KB: Migrate Virtual Machines (VMs) and VMkernel Adapters from Distributed Switch (vDS) to Standard Switch (vSS).

    1. Again, a new standard switch can be built using one of the physical adapters that passes the correct VLANs from the vDS. See the note below on Load Balancing information.

  3. Remove any remaining physical adapters from the vDS.

  4. Remove the host from the vDS. Click Home > Networking. Right click the vDS, click Add and Manage Hosts > Remove Hosts, and follow the wizard.
    NOTE: if there is an error about "Port-id is in use", follow this KB: The resource 'Port-ID' is in use error when removing a host from VDS.

  5. Add the host back to the vDS along with its vmkernel adapters, physical adapters, and VMs. After this step re-configure LACP/LAG again in both the virtual and physical the environment.

If vCenter Server or an ESX/ESXi host has not been restarted recently and the alert does not clear on its own, there may be other problems in the environment inhibiting the synchronization of vDS data. If the problem persists, ensure there are no intermittent network connectivity issues with the ESX/ESXi host or to vCenter Server (and vice versa). For more information, see Intermittent network connectivity on ESXi or host disconnect issues.

The following articles are also related to the issue:

Additional Information

To be alerted when this document is updated, click the Subscribe to Article link in the Actions box.

Powered by Wolken Software