Description

This article applies to a scenario with RHOSP cluster and Contrail Networking 2011.L2. It is only applicable if a stack update is performed with the stable/train version parent ID 232535ec.

It does not apply when a new cluster is deployed using the same template version parent ID 232535ec. This version has the files "deployment/contrail/contrail-config-database.yaml" and "deployment/contrail/contrail-analytics-database.yaml" updated with "user: root" permissions.

Symptoms

After performing a stack update, when the command, 'contrail-status' is executed on config/analytics database, the node manager gets stuck in initializing state.

Example contrail status:
[root@lxvcw1jd1001 ~]# contrail-status
Pod    Service    Original Name          Original Version   State  Id      Status      
database cassandra   contrail-external-cassandra   ubi-queens-2011-372 running 7537f29a09c8 Up 3 days ago  
database nodemgr    contrail-nodemgr         ubi-queens-2011-372 running 4f313d00ec90 Up 6 weeks ago  
database provisioner  contrail-provisioner       ubi-queens-2011-372 running 97de3b99a638 Up 6 weeks ago  
database query-engine contrail-analytics-query-engine ubi-queens-2011-372 running c7964033aad8 Up 6 weeks ago  

== Contrail database ==
nodemgr: initializing (Cassandra state detected DOWN. Disk space for DB not retrievable. )
query-engine: active
cassandra: active


Also, in the podman logs for the node manager container references issues with the root user:

CRITICAL:root:Error: unable to find user root: no matching entries in passwd file
ERROR:contrail-database-nodemgr:Failed to get database usage: Result: 1
CRITICAL:root:Error: unable to find user root: no matching entries in passwd file
ERROR:contrail-database-nodemgr:Failed to get database usage: Result: 1


However, the root user is present in the passwd file within this container:

(database-nodemgr)[root@lxvcw1jd1001 /]$ cat /etc/passwd | grep root
root:x:0:0:root:/root:/bin/bash

Solution

Rolling Restart of nodemanager service on all 3 controllers at 1 min intervals resolves the issue and no errors are seen then.

Restart can be done as below:

[root@overcloud-contrailcontroller-0 heat-admin]# docker ps | grep -i database_nodemgr
971dd46b0fd4        192.168.24.1:8787/contrail/contrail-nodemgr:2011.L2.372                        "/entrypoint.sh /b..."   4 months ago        Up 2 months                             contrail_analytics_database_nodemgr
33136efe81eb        192.168.24.1:8787/contrail/contrail-nodemgr:2011.L2.372                        "/entrypoint.sh /b..."   4 months ago        Up 2 months                             contrail_config_database_nodemgr
[root@overcloud-contrailcontroller-0 heat-admin]# 


Identify the name of the nodemanager which has the issues. For example, if it is Analytics Database nodemanager, then restart the analytics nodemanager as below:

podman restart contrail_analytics_database_nodemgr

Modification History

.

Related Information

  • Contrail Networking: 2011.L2
  • Template Version: Merged into stable/train version parent ID 232535ec
  • URL: https://github.com/tungstenfabric/tf-tripleo-heat-templates/tree/stable/train
  • Commit ID : 76b2597cf5244019d06cc440eaeb444e201a09e8
  • URL: https://github.com/tungstenfabric/tf-tripleo-heat-templates/commit/76b2597cf5244019d06cc440eaeb444e201a09e8