The settings looked correct and in the events list I could see vCenter trying to send the email, DNS was fine and there were no firewall issues. Tracing on Exchange showed that the email never reached a CAS server. This meant I had to dig through the vCenter logs on the VCSA.
1. Enable the SSH (if this wasn’t done during deployment) and the Bash Shell (https://VCSA_IP:5480)
2. SSH to the VCSA and type ‘shell’ at the ‘command>’ prompt. This will open the bash shell.
3. Look at ‘messages.log’ in /var/log (‘cat /var/log/messages.log | grep sendmail’ should be sufficient)
In here I discovered something strange. My SMTP server was set to ‘scan.domain.com’ but vCenter was trying to use just ‘domain.com’ as the relay address, therefore the connection was timing out.
2015-10-14T16:50:39.137459+00:00 VCSA1 sendmail: t9EGZCgq010289: to=<email@example.com>, ctladdr=<root@VCSA1.domain.com> (0/0), delay=00:15:25, xdelay=00:00:00, mailer=esmtp, pri=210969, relay=domain.com., dsn=4.0.0, stat=Deferred: Connection timed out with domain.com.
4. At this point I logged into the vSphere Web Client and modified the SMTP address to an A record that doesn’t exist (madeup.domain.com). Back in the SSH session, I could now see entries such as below, which verified that the change took effect.
2015-10-14T16:57:38.355977+00:00 VCSA1 sendmail: t9EGU4Yx022319: to=<firstname.lastname@example.org>, ctladdr=<root@VCSA1.domain.com> (0/0), delay=00:27:32, xdelay=00:00:00, mailer=relay, pri=300970, relay=[madeup.domain.com], dsn=5.1.2, stat=Host unknown (Name server: [madeup.domain.com]: host not found)
5. Changing the SMTP server back to a valid A record (scan.domain.com) which seemed to do the trick.
2015-10-15T12:27:32.504943+01:00 VCSA1 sendmail: t9FBRVI1028752: to=<email@example.com>, ctladdr=<root@VCSA1.domain.com> (0/0), delay=00:00:01, xdelay=00:00:01, mailer=relay, pri=120873, relay=scan.domain.com. [192.168.1.10], dsn=2.0.0, stat=Sent (<201510151127.t9FBRV6E028721@VCSA1.domain.com> [InternalId=13351197] Queued mail for delivery)
While the fix at the end was pretty straight forward the troubleshooting steps are useful to know.