Copy link to clipboard
Copied
I have a Win2008/IIS/ACF9 box that weekly sends out 36,000 emails through a separate SmarterMail15 email server set up specifically to be the bulk email server for this and another box running Lucee. The other box can still send emails without a problem. Starting on about Monday (not sure exactly when) the ACF9 machine now sends emails at the rate of one every 15, 30 or 60 seconds, and many of them simply get moved over to the Undelivr folder. Like, most of them.
In case this was an issue of routers or network cards, I installed another free SM server locally and changed the IP number and port in some of the .cfmail files to the local IP 127.0.0.1:9888, with no improvement.
This is not my expertise. I need someone who understands how ACF9 deals with email, and understands how email is handled in general.
1. Why is email moving so slowly suddenly from this machine, when the other Lucee box has no issues?
2. Why are so many of the emails being sent winding up in the Undelivr folder?
3. Is there a setting hidden in the XML that needs to be changed? (See link, below.)
I am looking for direct help, today, now, and would want to have a phone call / TeamViewer session to work on it till it's fixed.
Please contact me directly through this site. Let me know what you charge, and when you can jump on. I'm on the east coast and am ready to go.
...
sdsinc_pmascari wrote
We, too, send out large batches of email through CF to an external mail server. Have been doing so for many years with little problem. I'll just throw out some troubleshooting items to consider....
Did you check the CF Administrator settings for mail?
Yes. All settings were fine. I tweaked them a little with no effect.
Make sure the CF Admin is set to spool the mail to the email server. That way the email server is doing most of the work. If it is spooling, do you have e
Copy link to clipboard
Copied
You are unlikely to get that kind of support here.
If emails are being undelivered they will usually tell you why. Check the mail.log file for any issues.
Copy link to clipboard
Copied
There used to be a user, here, named Charlie Arehart (I think carehart-eiDyI5) who would attempt to help via the forum, first, then offer services of his consulting company. I haven't seen him around the Adobe forums for quite some time. I think he (like many others) moved to slack.com.
haxtbh is right about the CF logs. They should indicate what is causing the slowdown in sending the emails. If not, then check the server logs to see if anything has been hogging the CPU or RAM on the server. Even a 40% spike on the CPU can slow things to a crawl. Or maybe it's virtual memory. If the server swap file isn't big enough, or if there's less than 20% of the HD space free, that can slow things down, too.
HTH,
^_^
Copy link to clipboard
Copied
Checking the logs shows many errors, and that the individual email was moved to the Undelivr folder.
Copy link to clipboard
Copied
I realize this post is now a few months old and since the last reply. I have an update, in case it may help someone in the future.
Mik did in fact reach out to me that day (Apr 20) and we spent about an hour trying to resolve things. I could not find the answer during that call, but along the way and at the end I wondered if it may be a firewall issue. He got back to me later that day to confirm that indeed that was the issue. He had to whitelist the mail server's IP address in his firewall, and mail started sending. (I didn't charge him for that hour since I had not really "solved" it but only pointed to that suspicion which he confirmed on his own.)
BTW, to Wolfshade's comment, I'll note that the reason I had not been active on these forums was that for at least several months (or a couple of years, it feels), the mail process for the forum was not working (for me and many forum users). I am an inveterate mail guy (...so no, definitely I had not "moved to slack"!) But because I was not getting emails from the forum, I was just not inclined to try to stay up (visiting the web interface daily, for instance). Fortunately, just last month this problem was finally fixed (by Adobe) so you may see others like me who were "rather quiet" perking back up.
Finally, yep, as he notes, I try to answer questions on the forum (I've responded to dozens the past couple of weeks), and only once in a while do I mention my consulting, if I sense it may be the sort of problem where it just seems it would be more effective, if it may be a complicated problem with lots of possible variables. Or if I sense the person may have an urgent need, which is why (I suspect) Wolfshade had mentioned that here. Sure, happy to help folks, whatever way gets them to their solution fastest!
Copy link to clipboard
Copied
We, too, send out large batches of email through CF to an external mail server. Have been doing so for many years with little problem. I'll just throw out some troubleshooting items to consider....
Did you check the CF Administrator settings for mail?
Make sure the CF Admin is set to spool the mail to the email server. That way the email server is doing most of the work. If it is spooling, do you have enough hard drive space (or memory) to spool your large batch of message.
Is the timeout waiting for the mail server too low? Do you have the Maintain connection to mail server checked?
Is the mail server, itself, too slow or being bogged down by some other process?
Copy link to clipboard
Copied
sdsinc_pmascari wrote
We, too, send out large batches of email through CF to an external mail server. Have been doing so for many years with little problem. I'll just throw out some troubleshooting items to consider....
Did you check the CF Administrator settings for mail?
Yes. All settings were fine. I tweaked them a little with no effect.
Make sure the CF Admin is set to spool the mail to the email server. That way the email server is doing most of the work. If it is spooling, do you have enough hard drive space (or memory) to spool your large batch of message.
We do spool, and it goes to the /mail/spool folder. Yes, we have plenty of hard drive space. This was the folder that the mail was hung up in.
Is the timeout waiting for the mail server too low? Do you have the Maintain connection to mail server checked?
Turns out the mail server IP/ports needed to be added to the firewall. Processing was stopping with each NetBios query.
Is the mail server, itself, too slow or being bogged down by some other process?
Nope. It's a decent box and this is the only service running on it, other than DNS (as a secondary).
Copy link to clipboard
Copied
We used to perform similar mailing functions using ColdFusion 9, but gave up after even single messages would be processed by CFMail, but either never deliverable or mysteriously dropped. As a workaround, we set up SMTP service on each IIS server and deliver email to localhost and then have it rerouted to a non-local SMTP server for delivery. This seemed to be more stable, but we'd still encounter deliverability issues due to blacklisting & temporary greylisting.
My recommendation is to not use CFMail for bulk mailers. We switched to SparkPost's Transactional Email API and are sending up to 100,000 emails a month using their REST API. (SparkPost also has an SMTP connection, but ColdFusion caches IPs forever and doesn't honor TTL... so a service restart is required whenever their DNS is updated to point to a different mail server.) We're able to generate an JSON array of recipients (complete with personalized information) for each email. Most bulk mailers are sent within 10-15 minutes. Any complaints or unsubscribe requests are automatically suppressed, logged and can be reported back to your ColdFusion server using a webhook. All of our clients have been pleased with the migration... especially during the last Christmas season.
Copy link to clipboard
Copied
@jamo
We tried the local SMTP thing as well with the same results.
Sparkpost sounds interesting. I need to handle bounces and spam reports better.
I swear, email is a serious p.i.t.a., from delivery, to DNS (and all that implies), to being reported as a spammer, and more. Such a major headache. I hate it.
Copy link to clipboard
Copied
The SparkPost dashboard provides a filterable engagement graph that highlights TARGETED, ACCEPTED, UNIQUE CONFIRMED OPENS and UNIQUE CLICKS. I also like that they have API endpoints for just about anything that you want to do. (We were able to get everything to work using ColdFusion 9.)
We still use ColdFusion internally to accept messages to be sent... but if SparkPost is the configured sender, instead of looping over every email address using CFMail, we generate a single JSON packet and post it to the SparkPost API. This allows us to maintain a log of messages sent and enables our clients to schedule messages to be sent in the future. SparkPost can also send messages in the future, but we were able to add more customizeable rules (ie, edit-and-resend, import new list of recipients and send, test message, etc)
We initially used Mandrill, but they dumped all of their free plans and didn't give their developer-base much warning. (We are very glad to have been forced to find another provider.) Here's the list of transactional email providers that we evaluated. Be sure to compare the features that you need.
http://socialcompare.com/en/comparison/transactional-emailing-providers-mailjet-sendgrid-critsend
Copy link to clipboard
Copied
I had this same issue with sending out bulk emails when using CF8 and CF9. I was sending directly from CF to Sendgrid, and would sometimes get backlogs where the emails were taking 2+ hours to leave my machine.
I adjusted my coldfusion instance to not queue the emails, but send directly to my own machine, where I setup IIS6 to handle as an SMTP relay. This way the emails were immediately outside of CF, and IIS was very quick to relay the emails to Sendgrid with their authorization via SMTP.
After this change, my emails went from having a backlog that could never be cleared, to all emails going out instantly.