Jira Mail Queue sending mail incredibly late, if at all.

[Ryan Gavin]

Our Jira Mail Queue is getting backed up and not sending messages until two days or so after submission. We use mail any time a comment is made on an issue to keep users alerted. Flushing the mail queue and bypassing currently sending mail is of no use, either.

Insight on the issue would be greatly appreciated. Please let me know what information we can provide to better help the situation.

Thank you.

[Ryan Gavin]

Thanks for the input, Alex. We've been using Jira Server.

We've recently migrated our environment from on prem systems to AWS. Everything was migrated over with all settings as close as possible. The old environment worked normally. It seemed like all was relatively well in the beginning but it has now gotten worse. If anything the environment is more polished and our hardware is better so now so we're even more curious.

This has been lingering on now for over a week, so we don't think it's getting backed up from a single process. Any other insight or ideas?

[karthikeyan santhanam]

Ryan, do you have any luck on the JIRA mail going out with no delay with your JIRA being in AWS? Even we are experiencing the same delay in mail queue, our JIRA 7.0 in AWS. 

We also made the change to 1 * * * * ? for Mail Queue Service in Admin->System->Services. 

[Ryan Gavin]

Hi Karthikeyan,

We have resolved the issue. It was a combination of DNS settings, changing the mail queue service refresh interval, as you mentioned, and running an sql statement on system start that made it so jira won't check IPV6 before IPV4.

Thank you for reaching out, hopefully this gives insight to your troubles as well.

[karthikeyan santhanam]

Thank you so much for the input Ryan and that was so quick :)

Can you please give me some more details on the sql statement to run on system start and DNS settings changes? 

[Ryan Gavin]

No problem at all. Here's a link to the IPV6 fix. Unfortunately, the dns problems were resolved by our devops team and I am unaware to what the fix was exactly.

https://confluence.atlassian.com/jirakb/unable-to-connect-to-mail-server-due-to-ipv6-issues-on-jvm-637731067.html

Hopefully this helps! I'd recommend submitting a helpticket with Atlassian support as they will be able to look through your support logs, as they did in our case.

Good luck!

[Alex Christensen]

Are you using JIRA Server or Cloud?

I've seen this issue happen on JIRA Server if there's a lot of activity that's configured to have a lot of notifications go out to a lot of people for a lot of events. For example, we had someone submit a bulk change for 10000+ issues (10 different bulk changes of 1000 issues each, at least) and they forgot to uncheck "Send mail for this operation" in the bulk change process. So, our system had to send out 10000+ emails to everyone configured in the notification scheme for that particular issue event. It took a day or two for the mail queue to catch up.

In addition to that, it's good to revisit your permission schemes to make sure there aren't any unnecessary people are getting notified. It probably also depends on your mail server, but I believe a mail server can only send so many messages at once or quickly enough to not get backed up like it did for you.