A ticket was created with the wrong (old) SystemID

Technical assistance
1

Infos:

  • Used Zammad version: 6.5.2
  • Used Zammad installation type: package
  • Operating system: Ubuntu 24.04.4
  • Browser + version: Firefox 147.0.4

I changed the SystemID from 36 to 37 after the restore process. This in my opinion was required as hardware failure left me with a month old backup, consequently new tickets for a while would have been assigned used ticket numbers.
After the restore process, new tickets were assigned numbers in the correct format starting with 37. Notification emails for the initial few tickets showed older ticket numbers starting with 36 but afterwards both the tickets and the notifications showed the correct numbers.
A day later, triggered by a customer’s email, Ticket#36794 was apparently created as confirmed by the auto reply and notification emails. However, this ticket isn’t accessible in the system and the link in the notification leads to an older ticket. This problem hasn’t repeated in the subsequent tickets but I am worried it may occur infrequently in the future and also curious about the reason. I have stopped and stared the instance in hopes of “flushing” the older SystemID.
Would you recommend that I start with a completely new instance without the older data to avoid ever encountering the same problem, as the system is new enough for in prod for this to not matter in the long term.

Expected behavior:

  • Correct ticket number for an email ticket should have been Ticket#37803. And the ticket should be visible and accessible in the system

Actual behavior:

  • Ticket was assigned Ticket#36794 and could not be seen in the system or through the link in the notification email.
2

I have faced this issue again with a new ticket. This time I received two email notifications for a single email sent by the customer. One had the ticket number in the series 36 and one started with the correct series 37. Although the link in the email with the Ticket number 36796 lead to an older ticket.

3

Sounds like you have two different instances runninh to be honest. Is the old instance really shut down?

4

Hello. Thank you for replying. The old instance cannot be running as I reinstalled the OS before the restore process. The old instance was a docker container while the new one is a package install.

5

Any clones / test systems maybe? The behavior you‘re describing is an absolute classic.

Unless you receive notification for existing tickets (before the restore) that have the old ticket number. Those aren‘t updated of course.

6

I think you are right. While there is no test instance on the server itself. Before restoring, I tested the whole process on a WSL instance and it may have been running both times I received duplicate notifications. I have stopped it now.
Thanks a lot. This may have stopped me from nuking the system and restarting again.