RabbitMQ logs are not being collected by Fluentd
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
kolla-ansible |
Fix Released
|
Medium
|
Lev Klimin | ||
Rocky |
Fix Released
|
Medium
|
Mark Goddard | ||
Stein |
Fix Released
|
Medium
|
Lev Klimin | ||
Train |
Fix Released
|
Medium
|
Lev Klimin | ||
Ussuri |
Fix Released
|
Medium
|
Lev Klimin |
Bug Description
The following report is based on a Centos deployment.
On Rocky:
* RabbitMQ log name defaults to `<email address hidden>`. The hostname is included by default because the log file name is not set in the Rabbit config. Fluentd is set to ingest the RabbitMQ log from `rabbit.log`, so it is never picked up.
* When the Fluentd config is fixed so that the log is ingested the pattern matching is broken, and Fluentd spits out parsing errors.
* We don't collect the SASL log
In Stein:
RabbitMQ is upgraded to 3.7.10 from 3.6.5
* The same issue with log file naming as in Rocky exists.
* The SASL log (which is not ingested in Rocky) becomes part of the main RabbitMQ log and the log format changes.
* An upgrade log now exists
* Crash logs now exist
The expected behaviour is at the very least Fluentd should ingest the main rabbitmq log so that we can alert on failures.
Changed in kolla-ansible: | |
assignee: | nobody → Doug Szumski (dszumski) |
Changed in kolla-ansible: | |
importance: | Undecided → Medium |
Fix proposed to branch: master /review. opendev. org/692976
Review: https:/