VMware Networking Community
amb1s1
Contributor
Contributor

Error when creating a Logical Switch - A specified parameter was not correct: spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort

I'm trying to create a logical switch and I get the following error:

Error creating network vxw-dvs-19864-virtualwire-8-sid-5003-LS3 - A specified parameter was not correct: spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort; nested exception is com.vmware.vim.binding.vmodl.fault.InvalidArgument:

invalidProperty = spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort

inherited from com.vmware.vim.binding.vmodl.fault.InvalidArgument: A specified parameter was not correct: spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort.

I see between the messages in the error a problem with the uplink port order. In my case I have two VTEP and the teaming is Load Balance - srcid.  Any help will be appreciated. Thanks

0 Kudos
7 Replies
cnrz
Expert
Expert

Does this error observed on NSX manager or ESXi Host, also are the ESXi hypervisors nested mode or bare-metal (non-nested)? There may be some restrictions or settings that need to be done for nested environment as wel as the Uplink Load Balancing. Also are there any other Logical Switches working on the NSX, if VTEPs are successfully deployed and hosts are prepared creation of additional LS should be successful. If this is the first logical switch, it may be helpful to check the Host and VXLAN Preparation.

0 Kudos
kmanne
Contributor
Contributor

I see the same error. I have lacp enabled on the DV switch and 2 LAG uplinks. I already have 3 logical switches created. This started happening after ugprade nsx to 6.2.4.

Error creating network vxw-dvs-106-virtualwire-12-sid-5003-tr - A specified parameter was not correct: spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort; nested exception is com.vmware.vim.binding.vmodl.fault.InvalidArgument:

invalidProperty = spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort

inherited from com.vmware.vim.binding.vmodl.fault.InvalidArgument: A specified parameter was not correct: spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort.

0 Kudos
cnrz
Expert
Expert

If the previously created switches before the upgrade to 6.2.4 continues to work, but new Logical Switchces can not be created, then creation of LS may be broubleshoot. During the Logical Switch creation, NSX Manager sends the necessary API commands to the Controllers and vCenter Server. Controllers create the LS, and vCenter server creates a Port Group for this logical switch.

The error shows that VNI 5003 is assigned to the switch by the NSX Manager, so NSX Manager logs about this Vxlan may be helpful:

  1. Are there any error messages about the creation of the LS on NSX Manager:

show manager log reverse

2. show logical-switch list all --> If NSX Manager has created the LS, this sould be listed with the Name and the VNI



So it may be possible that controllers know about this LS.   To check if the Controllers know about this LS is it possible to send the output of this command: (from the Controller CLI)

2.  show control-cluster logical-switches vni 5003

and previously created LS as show control-cluster logical-switches vni 5001

If thecontroller part is successful, then most probably vCenter cannot create the Port Group. It can be concentrated on vCenter and ESXi hosts.

3.  From one ESXi host:


esxcli network vswitch dvs vmware vxlan network list --vds-name Name_of_the-dVS_for_Vxlans--> This command should list VNI 5003 along with previously created VNIs. If 5003 is not listed then vCenter cannotcreate  publish the Port Group to ESXi hosts.

(This link may be helpful for other commands)

https://vcdx133.com/2015/05/14/nsx-v-troubleshooting-deep-dive/

LACP Teaming Policy  does not support Multiple vTEPS, is it also possible to check single vtep is used for ESXi hosts?

http://www.routetocloud.com/2014/05/nsx-and-teaming-policy/

0 Kudos
kmanne
Contributor
Contributor

Previously created LS work. But creation of new LS is failing.  I have attached the nsx mgr log when the LS creation fails.

The other commands dont give any output as the LS is not created on nsx mgr. -

SDC-NSX-mgr>  show logical-switch list all

NAME                 UUID                                 VNI        Trans Zone Name      Trans Zone ID

SDC-NSX-mgr>

nsx-controller # show control-cluster logical-switches vni 5000

Error: Not found

[root@localhost:~] esxcli network vswitch dvs vmware vxlan network list --vds-name=Compute-DVS

[root@localhost:~]

[root@localhost:~] esxcli network vswitch dvs vmware lacp config get

DVS Name     LAG Name     LAG ID  NICs           Enabled  Mode     Load balance

-----------  --------  ---------  -------------  -------  -------  --------------------------

Compute-DVS  N5k-LAG   120273208  vmnic4,vmnic5     true  Passive  Src and dst ip, port, vlan

vxlan teaming policy is failover mode and uses 1 vtep.

0 Kudos
cnrz
Expert
Expert

NSX Manager does not see any Logical Switches, it is normal for the new LS, but the other Logical switches also does not list.  Are there any switches listed on the GUI? If there are configured Logical Switchces, but NSX Manager can't see them from the CLI this may be related to NSX manager DB or a similar issue.

For the VTEP Vxlan, if failover is chosen to be Teaming Method, there is no need for LACP. Can there be any problem with the LACP configuration on the ESXi or Physical switch side? Are there any other port groups on the Compute-DVS that use the same uplink, if they are working properly this may mean LACP configuration is working properly.

Also, is it possible to check if the same is occuring again with VXLAN VTEP Configuration with choosing LACP Active instead of Failover? Since this is done at the Cluster level, it may effect the other hosts, but for consistent configuration for dVS and VXLAN on NSX this configuration may be recommended if there are any differences about the error occuring. Since the error seems to be about Teaming Uplink Policy,   LACP configuration may be important.

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=20342...

Configuring LACP on an Uplink Port Group

Note: All port groups using the Uplink Port Group enabled with LACP must have the load balancing policy set to IP hash load balancing, network failure detection policy set to link status only, and all uplinks set to active.

0 Kudos
manmeet90
Contributor
Contributor

As per my understanding there are some changes on the distributed switch recently from its initial configuration.

It can be anything from the Name of the uplinks to Teaming and failover policy. I suggest rolling back any changes made on DVS that may differ to what NSX has in its database.

You can follow the steps in the KB - https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=21102...

and check for what is current config that NSX has in its database, compare this with details you see in vcenter server DVS config and correct any miss configuration.

I did check the logs attached.

2016-09-15 22:04:41.265 GMT  INFO http-nio-127.0.0.1-7441-exec-2 DistributedVirtualSwitchVcOperationsImpl:152 - Removing the pg vxw-dvs-106-virtualwire-25-sid-5000-test from the dvs map for switch id dvs-106

2016-09-15 22:04:41.266 GMT ERROR http-nio-127.0.0.1-7441-exec-2 VcDriver:302 - Error adding portgroup : A specified parameter was not correct: spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort; nested exception is com.vmware.vim.binding.vmodl.fault.InvalidArgument:

invalidProperty = spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort

inherited from com.vmware.vim.binding.vmodl.fault.InvalidArgument: A specified parameter was not correct: spec.uplinkTeamingPolicy.uplinkPortOrder.activeUplinkPort

2016-09-15 22:04:41.267 GMT ERROR http-nio-127.0.0.1-7441-exec-2 VirtualWireServiceImpl:1258 - Could not create backing for virtual wire: test

Are we really using LACP in the environment? for the VXLAN vteps, if yes please check and confirm the teaming and failover under Network & Security > Installation > Configure VXLAN tab and change to IP hash.

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=20570...


Last possible reason for this issue the stale entries for this Logical switch in the NSX DB, that needs to be cleared manually.

0 Kudos
vwic
Contributor
Contributor

I am observing the same symptom with my NSX on the same version.

The only change was when I removed an unused dvUplink (dvUplink3) from the dvSwitch and the creation of new LS started to fail. dvUplink1 is active, dvUplink2 is standby.

Re-adding dvUplink3 as an unused uplink did not resolve the issue.

0 Kudos