I encountered the same problem and investigated a bit further...
The SparkPost extension does not directly use mcrypt but it does use CRM_Utils_Crypt which makes use of mcrypt if it is available.
If you upgrade to php 7.2 without mcrypt and you already have an API key saved, the result is that the extension cannot decrypt the key. That's why the ...
PHP 7.2 does not have the mcrypt module, and the Sparkpost extension needs that module. If you can install php7.2-mcrypt as a pecl module Sparkpost will work. If not, I would move to php 7.1 for the time being.
I've used them both. CiviDesk's version wins hands down on ease of installation as you just install and plug in your API from SparkPost. Also if I understand it, sending via the API is preferred.
PESC's version requires installation, update SMTP settings with those provided by SparkPost, and then configuring a Scheduled Job with API and other configuration ...
See if https://github.com/cividesk/com.cividesk.email.sparkpost/pull/77 fixes your problem. It's been merged but there has not been a release of the sparkpost extension since then so either pull from git or apply the fix manually - it's only one line of code.
I found this on the github issue list: https://github.com/cividesk/com.cividesk.email.sparkpost/pull/58/commits/69a0bdb17bb520334da7f472a2d7670b44da8071
Applied the commit and it has resolved the problem.
CiviCRM has a utility class CRM_Utils_Crypt for encryption & decryption. If PHP Mcrypt is installed, it uses that plus base64 encoding. If not, it falls back to using just base64.
This class is used for encrypting some settings, including:
smtpPassword in core 4.6.30
sparkpost_apiKey in com.cividesk.email.sparkpost
PHP Mcrypt was not initially ...
I have two outside ideas that 'might' be helpful:
Is it possible Activities have been made for the emails that should have gone out? Emails from the Contact record will almost certainly have created an activity, and it's a commonly set option for emails sent via Scheduled reminders. If an activity has been made it will probably contain the email.
Do you ...
Your question does not say why you can't use DNS?
Google Apps will I think let you configure the abuse@ and postmaster@ addresses as aliases (Google refers to these as "groups"). Google will still process those addresses as well, but I think you can configure them so you can field the Sparkpost setup email.
Here is Sparkpost docs on that: https://www....
I use the following settings, basically the default:
Mailer batch limit: 0
Mailer job size: 0
Mailer throttle time: 0
Database Update Frequency: 10
I use a long-running cron, so it's really just one process running at 100% while sending emails. I don't want to impact too much the rest of the CRM, so I only run one thread/job at the time.
One easy tweak ...
"One thing that puts us off a dedicated IP is the extra admin that would be involved in keeping the IP off blacklists. "
That's one of Sparkpost's responsibilities (although you should also monitor, just in case, I particularly like http://multirbl.valli.org/lookup/).
"If you are on a dedicated IP, how many emails per month are you sending? Has ...
Same advise here, we have all our customers on a few private IPs as this is the only way to not end-up been blacklisted. If you are sending with the default provider's service, tens of thousands of other domains also are, and nobody controls what they are sending - odds are that there will be at least one 'bad apple' that will get the IP blacklisted and ...
It seems like you were missing the "Unsupervised" rule in CiviCRM (from Contacts > Find and Merge Duplicate Contacts).
The default rule, by email, was set as "General".
I changed it back to "Unsupervised".
It's rather odd that CiviCRM allows this, I will see if we can add a check against it.
The problem was caused by the index.php in the links. I fixed by changing the WordPress Settings --> Permalinks
It was set for Custom Structure and when I changed it to Day and Name it fixed the problem. Now if I could just get the opt-out working I'd be happy...
I have had similar issues recently, after I have manually ran some random scheduled jobs forcefully.
Later I have changed the time of the scheduled job, specially for the ones that runs daily, to the default (12:00).
Since then my issue resolved. I would suggest try that.
Two weeks ago I would suggest Mailgun but the plugin supplied for webhook settings seems to be updated to work for new CiviCRM versions.
But if you somehow make it work (and so integrate the service with Civi successfully) the service itself is quite successful for most of the mail providers.
Here comes the last two weeks... Yahoo (and AOL) domains keep ...
OK, the answers here finally sunk in. The key is Aidan's answer in my mind, to simply use the same PHP version in your cron job as in your CiviCRM setup. Thanks, @Aidan, this is the 2nd (unrelated) problem you solved for me today! In our case, we run CiviCRM 5.10.3 under PHP 7.2.11 with CMS WordPress 5.2.1. After updating our Sparkpost API Key at Sparkpost ...
How many emails are you trying to send? mailer cron=8 is really if you want to send a lot (I would not consider putting something else than 1 if less than 100k emails sent).
As for the job size at 1000. you are on the lower side, and would probably put at least an order of magnitude higher.
With more "normal" settings, you should get a better behaviour.
This appears to be a rather old bug. I opened an issue here and also a pull request.
The fix will be shipped with CiviCRM 5.12. Meanwhile, it's a 1 line patch that can be applied on top of CiviCRM 5.11.
The bug was specific to sending Thank You letters by email with the email address of the contact (even if it's an allowed domain, it was not specific to ...
Getting your own IP with SparkPost doesn't lessen their activity or responsibility under their contract to provide whitelisting for you. It probably makes it easier for them to respond quickly and effectively when it does happen. We don't monitor blacklisting ourselves because these services will notify us immediately if we appear on any blacklist.
One of ...
There's a forum discussion here which I think helps towards the answer. If you look at the table civicrm_mailing_bounce_type there's a threshold of 3 for bounce type 'Syntax'. I just encountered a bounce for type 'Host' that didn't place the email address on hold, so was also concerned as to why. The answer is in this table.
CiviRules provides a good foundation for say 'if trigger X happens, then do Action Y'.
I have no idea if the option you need is there but if it isn't it might be a good place to add it.
In other words, create a rule that says, if a Group is created then add contact X to it.
All you need to do is follow the documentation, as suggested by Karin.
When you save the API key in the setup screen, the extension will make API calls to SparkPost and create a callback. Whenever a bounce occurs, SparkPost triggers this callback and CiviCRM can then process the bounce event accordingly.
As also indicated in the documentation, only ...
Start here for background reading:
I found all I had to do was to just follow the Documentation: https://docs.civicrm.org/sparkpost/en/latest/install/
Bounces are processed through a callback (no need for an email account dedicated to bounce processing), but CiviCRM only processes ...