Hi, am a newbie Nutanix customer and have had Pulse issues since day 1.
We have new x2 8-node Nutanix clusters running ESXi6.5. These are defined as a metro cluster and all is working fine....apart from Pulse. The relevant knowledge base articles have been absorbed and understood. SMTP was initially set up per cluster and the self tests worked. However, the support portal shows Pulse is not communicating to Nutanix.
As the nodes exist on a private vlan, we assumed that each cluster would communicate through Prism Central which is hosted on a public vlan with access to the web etc. In desperation I used the proxy method from each local cluster and pointed these at PC which resulted in the correct connection being verified via nsc02.nutanix.net
so it appears to be communicating - but goes up and down regularly.
Running the following from the PC CVM results in a positive response:
curl -k https://insights.nutanix.com:443
Another test run:
Just another test from a local Prism server:
ncli> pulse-config ls
Status : Enabled
Enable Default Nutanix... : true
Default Nutanix Email : firstname.lastname@example.org
Email Contacts : x
Verbosity : BASIC_COREDUMP
SMTP Tunnel Status : success
HTTP Proxy : xxx.xxx.xxx.xxx
Service Center : nsc02.nutanix.net [nsc02.nutanix.net]
Tunnel Connected Since : Fri Nov 23 11:30:04 GMT 2018
Probably a really dumb question but if a proxy is defined, does this bypass SMTP - because today we received the following email notification:
Critical : Failed to send email from Controller VM xxx.xxx.xxx.xxx via xxx.xxx.xxx.xxx due to error: Connection refused.
Impact : Nutanix support cannot proactively monitor cluster health and offer assistance before issues become acute.
Cause : The cluster cannot send Pulse and alert emails to Nutanix support with the current configuration.
Resolution : Ensure an SMTP server is configured and reachable from the cluster. If no SMTP server is configured, ensure that the firewalls allow the cluster to send email messages through the Nutanix SMTP servers (nsc01.nutanix.net and nsc02.nutanix.net) on port 80.
Fully appreciate this may be user error or solely network\firewall related, but have been advised that the necessary security should not prevent this feature from working.
We are delayed going into production because of this so just wondered if you community could advise on anything further. We have engaged with Nutanix support but they can only suggest so much and in this case, I don't feel its a Nutanix issue.