New Nginx Manager and Certificate Errors

jhawkins003

Active Member
Jun 24, 2014
36
7
58
cPanel Access Level
Root Administrator
Greetings! Before I open a ticket I want to see if anyone else has run into this issue or has ideas.

We moved from Engintron to the new cPanel Nginx Manager for testing on our dev server this week and we have noticed that every morning all sites running Nginx fail to load due to security/certificate configuration issues. I assume this is tied to the nightly update runs. We also noticed this issue after installing Nginx Manager for the first time.

If we manually toggle a domain off/on in the Nginx Manager the site will come back online and appears to be configured correctly. Interestingly, if we globally disable/enable Nginx - that does not solve the problem. Only toggling each domain in the account listing interface within Nginx Manager seems to rebuild the config correctly.

Is anyone else seeing something similar? Are there any configuration updates we should be on the lookout for?
 
Last edited by a moderator:

jhawkins003

Active Member
Jun 24, 2014
36
7
58
cPanel Access Level
Root Administrator
Hey there! I haven't heard about this just yet on my end, so for such a new feature a ticket might be best so we can investigate that directly on the system.
Will do! A ticket has just been submitted and I'll post back to the thread with any updates that might be helpful to others.
 

jhawkins003

Active Member
Jun 24, 2014
36
7
58
cPanel Access Level
Root Administrator
We found a solution! An account that started with the letter "a" had a configuration bug. From the ticket - with 'alpha' substituted for the actual user ID:

...There was an empty line in the alpha user's /var/cpanel/users/alpha file:
DNS1=

Removing this and rebuilding the /etc/userdomains and /etc/userdatadomains resolved the issue, and the EA-Nginx configurations rebuilt.
Since the error occurred on the first domain Nginx tried to parse, none of the subsequent domains were properly handled by the Nginx config tool. When we manually toggled every domain, they were configured individually - which is why that work-around functioned for us.

So a few final thoughts on this experience:
  • cPanel support was professional and excellent as always - if a bit on the slow side - but this was not an urgent issue, so no objections :)
  • Nginx Manager reported an error to us when it attempted to configure itself but did not provide output as to what the error actually was. If we had known it was related to the 'alpha' account, we probably would have tried re-provisioning that account which likely would have built a clean configuration for it and solved the Nginx Manager issue.
  • A single account experiencing a bug seems like something the Nginx Manager configurator should error trap for so that all subsequent domains it attempts to manage don't break.
 
  • Like
Reactions: cPRex