Incident Summary
Bounce messages were not being populated in the message central database, which in turn meant that the bounce messages were not making it to the Advantage DB.
Leadup
No new deploys or configuration changes were deployed on the systems.
Fault
There was an extra process that was started on the MTA boxes. This process is the default out of the box log processing script. For the MC cluster, there are 2 other custom log processing scripts that populate the DB. When the default process was started it caused a conflict with the custom scripts.
Impact
Bounce data was not populated to the MC or Advantage database.
Detection
While investigating a similar issue, a member of the Deliverability team noticed that the bounce data was missing from the Advantage database.
Response
Members of the Deliverability, Command Center, Message Central, and Email Ninjas team responded to the incident as soon as it was identified.
Recovery
The issue was resolved by stopping the default process on all the MTA nodes.
The team then removed the default process from /etc/init.d stop it will not be started again.
Timeline
Root Cause
There was an extra process that was started on the MTA boxes. This process is the default out of the box log processing script. For the MC cluster, there are 2 other custom log processing scripts that populate the DB. When the default process was started it caused a conflict with the custom scripts.
Recurrence
This had never happened before.
Lessons Learned
We learned that we need to remove the default process to prevent it from being started unintentionally. The team worked well together to get the issue resolved, which further confirmed that the incident management process is effective.
Corrective actions
The Deliverability team removed the default process from /etc/init.d so that it will not be started again.