This update fixes a bug that caused the `metrics_qdr` service to appear in the HAProxy configuration of an Service Telemetry Framework (STF) deployment on a distributed compute node (DCN) node.
Description of problem:
RHOSP 16.2 DCN with Storage(Ceph), integrating with STF
3 x DistributedComputeHCI
7 x DistributedComputeHCIScaleOut
metrics_qdr service configured in haproxy.cfg on DistributedComputeHCIScaleOut and listening 0.0.0.0:5666, causing metrics_qdr container failed with bind address error.
Here is the configuration in haproxy.cfg on DistributedComputeHCIScaleOut node.
============================
~
listen metrics_qdr
bind :5666 transparent
option tcp-check
option tcplog
tcp-check connect port 5666
server controller01~
server controller02~
server controller03~
============================
From /usr/share/openstack-tripleo-heat-templates/deployment/haproxy/haproxy-edge-container-puppet.yaml, seems only glance-api services should be set in haproxy.cfg, but somehow metric_qdr service also configured in haproxy.cfg on the DistributedComputeHCIScaleOut node.
Version-Release number of selected component (if applicable):
RHOSP 16.2
How reproducible:
Steps to Reproduce:
1. disable public virtual IP
2. Deploy RHOSP 16.2 integrating with STF in DCN environment
Actual results:
metrics_qdr service configured in haproxy.cfg on DistributedComputeHCIScaleOut and listening 0.0.0.0:5666, causing metrics_qdr container failed with bind address error.
Expected results:
metrics_qdr container should be started without problem
metrics_qdr should not be configured in haproxy.cfg on DistributedComputeHCIScaleOut node.
Additional info:
Hi Meiyan. Thanks for creating BZ on behalf of me.
I included Oliver Walsh who helped with providing workaround for this issue.
parameter_defaults:
DistributedComputeScaleOutExtraConfig:
tripleo::haproxy::metrics_qdr: false
DistributedComputeHCIScaleOutExtraConfig:
tripleo::haproxy::metrics_qdr: false
According to his insight, the issue is in haproxy-edge-container-puppet.yaml shown the link below
>> https://github.com/openstack/tripleo-heat-templates/blob/stable/train/deployment/haproxy/haproxy-edge-container-puppet.yaml#L109.
I really appreciate all of your help and hope this issue can be permanently resolved in the next version,
meanwhile I will go with the workaround that Oliver shared.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory (Moderate: Red Hat OpenStack Platform 16.2 (openstack-tripleo-heat-templates) security update), and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHSA-2022:0995
Comment 22Red Hat Bugzilla
2023-09-18 04:30:00 UTC
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 120 days
Description of problem: RHOSP 16.2 DCN with Storage(Ceph), integrating with STF 3 x DistributedComputeHCI 7 x DistributedComputeHCIScaleOut metrics_qdr service configured in haproxy.cfg on DistributedComputeHCIScaleOut and listening 0.0.0.0:5666, causing metrics_qdr container failed with bind address error. Here is the configuration in haproxy.cfg on DistributedComputeHCIScaleOut node. ============================ ~ listen metrics_qdr bind :5666 transparent option tcp-check option tcplog tcp-check connect port 5666 server controller01~ server controller02~ server controller03~ ============================ From /usr/share/openstack-tripleo-heat-templates/deployment/haproxy/haproxy-edge-container-puppet.yaml, seems only glance-api services should be set in haproxy.cfg, but somehow metric_qdr service also configured in haproxy.cfg on the DistributedComputeHCIScaleOut node. Version-Release number of selected component (if applicable): RHOSP 16.2 How reproducible: Steps to Reproduce: 1. disable public virtual IP 2. Deploy RHOSP 16.2 integrating with STF in DCN environment Actual results: metrics_qdr service configured in haproxy.cfg on DistributedComputeHCIScaleOut and listening 0.0.0.0:5666, causing metrics_qdr container failed with bind address error. Expected results: metrics_qdr container should be started without problem metrics_qdr should not be configured in haproxy.cfg on DistributedComputeHCIScaleOut node. Additional info: