Service apache2 not started

Bug #1935990 reported by Hybrid512
10
This bug affects 2 people
Affects Status Importance Assigned to Milestone
OpenStack Dashboard Charm
Confirmed
Undecided
Unassigned

Bug Description

Hi,

I have an issue with my Openstack deployment, from time to time, in a "haclustered" configuration, some openstack-dashboard units fails to start with a message saying the apache2 service is not running.

Fixing the issue is pretty easy, either start the service by hand or reboot the unit.

There is no error that I can spot, it just seems that restarting service during deployment, probably after configuring the certificates relation to vault, the service fails to start for unknown reason and the charm doesn't try to restart it from time to time.

Would it be possible to make that service status check a bit more solid in order to avoid manual steps during deployment ?

Here is my deployment ocnfiguration :

===========================================================================
series: focal
applications:

  openstack-dashboard:
    charm: cs:openstack-dashboard-314
    channel: stable
    num_units: 3
    to:
    - lxd:3
    - lxd:4
    - lxd:5
    options:
      enforce-ssl: false
      offline-compression: "yes"
      openstack-origin: distro
      secret: iS6aV6dY1kC6gS1j
      use-syslog: false
      vip: 192.168.111.220
      worker-multiplier: 0.25

  openstack-dashboard-hacluster:
    charm: cs:hacluster-76
    channel: stable
    options:
      cluster_count: 3

  openstack-dashboard-mysql-router:
    charm: cs:mysql-router-10
    channel: stable

relations:
- - openstack-dashboard:ha
  - openstack-dashboard-hacluster:ha
- - openstack-dashboard:identity-service
  - keystone:identity-service
- - openstack-dashboard:shared-db
  - openstack-dashboard-mysql-router:shared-db
- - openstack-dashboard-mysql-router:db-router
  - mysql-innodb-cluster:db-router
- - vault:certificates
  - openstack-dashboard:certificates
===========================================================================

Revision history for this message
Bas de Bruijne (basdbruijne) wrote :

We are seeing this with SQA too.

From the logs it seems that something has gone wrong with the certificates:

apachectl[103299]: SSLCertificateFile: file '/etc/apache2/ssl/horizon/cert_10.246.173.14' does not exist or is empty

Crashdumps can be found here:
https://oil-jenkins.canonical.com/artifacts/e71d99fb-68a6-443e-b5fd-51f42322e4ec/index.html

Revision history for this message
Marian Gasparovic (marosg) wrote :

We are still hitting this issue from time to time in SQA tests.

Most recent occurance
https://oil-jenkins.canonical.com/artifacts/64dc0c06-ffb2-4392-a7fb-acdca4d47950/index.html

Changed in charm-openstack-dashboard:
status: New → Confirmed
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.