During the last month I had two clusters, where the LCM process didn´t start because of a missing lcm_family attribute for some of the nodes.
Unfortunately I can not find the error screenshot from PRISM, but the LCM error message contained lcm_family.
On each Nutanix node you find a hardware_config.json at /etc/nutanix with a bunch of lines with hardware and sensor details.
One line has to look like this (the value may differs between different hardware generations and vendors):
"lcm_family": "smc_gen_10",
In both cases this line didn´t exist on all nodes in the cluster.
This can be easliy checked from any CVM with the following command:
allssh "cat /etc/nutanix/hardware_config.json | grep lcm"
For each CVM one entry should appear in the output. In the following example you see it was missing for one of them:
nutanix@NTNX-16SM57610364-A-CVM:172.20.16.77:~/data/logs$ allssh “cat /etc/nutanix/hardware_config.json | grep lcm”
================== 172.20.16.77 =================
“lcm_family”: “smc_gen_10”,
================== 172.20.16.78 =================
================== 172.20.16.79 =================
“lcm_family”: “smc_gen_10”,
================== 172.20.16.80 =================
“lcm_family”: “smc_gen_10”,
The quiet easy solution is to modify the hardware_config.json file and add the missing line. Make sure its on the correct place within the “hardware_attributes” section like you can see here:

Now just run LCM again to get started with your pending updates.
NOTE: Please use the commands above only if you know what you are doing and at your own risk. If you are uncertain, I would strongly recommend involving the Nutanix support.
Speak Your Mind