@scottalanmiller said in Which is better NodeBB or Discourse?:

What sending mechanism are you using? Local SMTP? A service? We use MailGun and haven't seen an issue yet.

The out of memory may have been an independent issue. We use Sparkpost (as mentioned in the other thread).

@scottalanmiller said in Bounced emails:

@djensen47 said in Bounced emails:

Any thoughts here? Anybody? Is there any user setting or database record that basically says, "do not send mail to this account?"

Yes, my users control this for themselves.

But if the email is bouncing; I don't want to leave it to the user who may never actually come back. I have a user base of 12k going back as far as 2003.

@baris Okay, I think I'll go with changing the settings. Even if it's bouncing, I need the old email address for … reasons.

@baris Thanks this is super helpful!

Is there a way to "invalidate" an email (e.g., when a new user has registered but has not confirmed their email)? Would that stop emails from being sent?

Any thoughts here? Anybody? Is there any user setting or database record that basically says, "do not send mail to this account?"

Setting it to more than 2 hours after the original time seemed to work.

I tried to change the digest hour in the ACP to +1 hour and bounced the server. Nothing seems to be happening. Thoughts?

The same digest hour is used for daily, weekly, and monthly, correct?

For the first time reader of all of this, it is important to note that despite the digest email issue, the team is super responsive and receptive to every other issue. Especially when it is a well defined problem, the team turns around very quickly. I've logged a ton of bugs and nearly all of them get fixed and nearly all within a week of reporting. Priorities are priorities and this might be occurrence problem.

One thing to add to the list is the lack of reliability of sending the digest email to a moderate or larger user base. I have 12,000 users and have successfully sent 7/20 digest emails.

To the teams credit, one issue related to this has been resolved. Thank you for that.

Today, however, NodeBB ran out of memory. It appears this while it was getting ready to send the digest. The server crashed and the digest was not sent. The smallest error happens related or unrelated to the digest and it just doesn't start nor finish. And there is no way to resume or resend to those who did not receive it. The solution? Wait a week or wait until the "underlying issue" is resolved.

Update, possibly irrelevant:
The team does not believe this to be an resolvable (nor important enough) issue so it will not be resolved. (That's an opinion and their decision and that's fine). If you're looking for reliable weekly digest, you need to seriously consider something else besides NodeBB. I don't want this to be true but it just isn't a priority for them. (Development teams have a lot of issues pulling them in different directions; so you have to set some sort of priorities and some things get left out, this just happens to be one of them.)

I don't know if Discourse is better on this point but I am on a few Discourse forums that definitely have 10s of thousands of members and I get the email every week.

How do I reset the cron job to run, temporarily at a different time?

My server crashed during? Just before? The digest went out so zero digests were sent. I'm so frustrated. See below.

I think NodeBB is great but I'm also super frustrated by the lack of reliability of the weekly digest to perform on a set of uses as high as 12,000.

There have been 20 Sundays since I moved to NodeBB and the digest has worked 7 times. Most of those are in the last 6 weeks. It honestly is getting me to look at Discourse to understand how their digest works and if it is reliable and considering the switch after only 5 months on NodeBB. Not exactly a success story.