CUNY Academic Commons

(the reply that's there I posted directly)

Just got four emails from this post -- https://commons.gc.cuny.edu/groups/digital-humanities-initiative/forum/topic/fwd-invitation-to-digital-paxton-launch-events-2/.

And then just got four rejection messages each for my earlier replies.

Hi there,

Unfortunately, your reply to the forum topic could not be posted because it appears you have already made the same reply.

Here is a copy of your reply:

"RBE."

We apologize for any inconvenience this may have caused.

So the problem is the remote POST response used to trigger RBE is no longer responding:
https://github.com/cuny-academic-commons/cac/blob/1.10.x/wp-content/plugins/bp-reply-by-email/includes/bp-rbe-functions.php#L1350

I'm guessing IT has put in some security restrictions on POST responses. For the moment, I have manually triggered RBE. I'll look into adding some code that doesn't use a POST response to trigger RBE on the RBE settings page.

Boone, think this is worth a follow up with Lihua?

Let me do some more debugging before we contact Lihua.

I've written some new logic so RBE can spawn new inbox checks from the RBE admin settings page.

This approach doesn't use a loopback POST response as before, but instead uses AJAX to do the initial ping.

The good thing is this approach will probably stop our problem with RBE duplicate posts from occurring since you can only spawn new inbox checks from the admin settings page and not on a pageload basis like before.

The problem is if the Commons goes down or if RBE isn't able to connect to the inbox during a scheduled ping, someone will need to manually go to the RBE admin settings page and hit the "Connect" button.

To remedy this, I can write a task hooked to WP cron (perhaps run hourly) to check if RBE is connected. If RBE isn't connected, it can email some of us to go to the RBE settings page and re-connect.

FYI: This code is live on the Commons, but I haven't committed it to the CAC repo yet as I'm still monitoring things, but initial logs look good on my end.

It might be good to ping Lihua just to see what has changed exactly or to see if something like fail2ban is blocking the loopback POST response.