Thursday, April 18, 2019

NSX Host Preparation Failure reasons



 In the preparing steps of VMware NSX deployment, after NSX Manager and also controller nodes deploying (at least 3 nodes or other odd number of controllers) you need to do "Host Preparation" procedures in the ESXi clusters. Sometimes after execute "install" on the specified cluster, you may encounter some problems that are related to communication channel health, So the NSX Deployment phases will stop in this step. Today I want to review the cause of these problems:
1. Before executing any changes, please check the version of your vSphere suite (ESXi hosts belong to the cluster and vCenter server) and also the NSX Manager appliance version and review the compatibility of these products.
2. Check Lookup Service of PSC and NSX Manager connection. it can be related to the time settings that are not-synchronized or lack of same time value between these two servers. 
Also, you should Check port 443, maybe it has been blocked by a firewall. you need to check the certificates that have been used by vCenter and NSX manager, maybe they are self-signed Certificates and can cause trouble in SSL connectivity. each of them can be the cause of vCenter~NSX communication problems.
3. It's recommended to design and prepare vSphere Distributed Switches before deploying NSX in the SDDC, especially for preparing of VXLAN TCP/IP stack (and also VTEP state) in the SDN environment. So you need to attach all hosts of the cluster to the VDS and then deploy NSX on their cluster. Here is an important point you should remember: VDS and Cluster are not forced to have 1:1 relations (a VDS can be deployed to more than one cluster and a single cluster can have communication with multiple VDS) but when we talk about the NSX, it's better to have exactly a 1:1 relation between VDS and Cluster.

4. Check DNS settings of the ESXi management network stack (VMkernel) because it's may not do name resolution on Network Fabric for downloading VIB packages from the NSX Manager server.



5. Check the VXLAN configuration: IP Pools and vmkNIC status, try "resolve" option in the Host preparation and let the NSX Manager's Agents (Firewall and Control Plane) be get UP.



At last, if you change one of these settings, review every single operation that has been logged in the NSX Audit Logs and vCenter Events tabs.

No comments:

Post a Comment

I will start a new journey soon ...