I wanted to reach out in hopes someone may have ran into my particular issue. I currently have a fully functional VxRail "60" appliance with "4" nodes running smoothly with no issues. I wanted to expand this appliance with another appliance that was being utilized for testing. I ran a system build on the new appliance in order to bring it up to version 4.0.301 (same version as my existing one). After the rebuild, I started node "1" in hopes the node will be recognized by the existing VxRail Manager, however this didn't happen. I tried to restart loudmouth on both the new node and VxRail manager to no avail. I IP'd the new node so I can access its GUI and noticed that the initial configuration VxRail Manger VM was running, I shut it down restarted loudmouth but still nothing. I then realized that the first host on the new appliance was built as a primary node (default value), so I ran a reset on it and set it to a secondary node. After, restarting the new node again the existing VxRail Manager still was unable to see the new node.
I am unsure as to why the VxRail Manager Server is unable to identify the new node for expansion. Now here is the weird part, I ran "/usr/lib/vmware-loudmouth/bin/loudmouthc query | sed ‘s/\,/\n/g’" on the VxRail Manager and it identified node 2 on the existing appliance but not the new node. It should have not seen the existing node correct since it was already part of the appliance that has been deployed already and cluster built? Am I missing something, can this be why the new node is not being identified for expansion?
you can check vlan setting and muticast snooping setting for switch that it is connected.
and check loudmouth.log on vxrail manager and check it detect new node.
if not you can take network capture to investigate
The appliance that is being added, needs to be in a factory fresh state. If it is already or previously configured, you cannot add it to an existing cluster.
In order to make the "new" nodes factory fresh, you will need to run a reset. There are procedures in the VxRail SolVe desktop. If you are not an EMC employee or Service Enabled Partner, please reach out to your local Dell EMC VxRail account team for further assistance as these procedures are restricted.
Hope this helps
Yes both appliances are connected to the same switches. The first VxRail connected with no issues, I was able to run the initial configuration and it identified all four nodes on the appliance. However, the second appliance is not being detected by the first one.
The VxRail that is being added was configured with the matching version of the first VxRail. In this case 4.0.3 buidl server was used to build he second VxRail before trying to add it the the existing cluster. I even ran the reset script on all nodes and still the first VxRail is unable to identify the second VxRail. I went as far as to move all my VM's to a separate cluster and ran the build server on both VxRails and tried again. However, again I have no communication issues with the first VxRail since I was able to successfully run the initial configuration with no issues. When I turned on the first node of the second VxRail it was unable to be identified by the first VxRail.
I was able to verify the MLD snooping was working correctly on the switches. However, I am still perplex as to why the first VxRail still identifies the second node as not being provisioned. I am going to reset the first one again and configure in order to check logs if this happens again.
Does first cluster receive loudmouth from second cluster?
you can check it by looking into /var/log/vmware/loudmouth in vxrail manager.
If not, check network config