Bug #16215

Updates to pulp breaks httpd on capsule

Added by Dylan Baars almost 6 years ago. Updated almost 4 years ago.

Foreman Proxy Content
Target version:
Bugzilla link:
Pull request:
Fixed in Releases:
Found in Releases:
Red Hat JIRA:


After running "yum -y update" on our katello capsule (version 1.11.4 reported in the GUI) these updates were applied

Aug 18 19:22:41 Updated: python-pulp-common-2.8.6-1.el7.noarch
Aug 18 19:22:41 Updated: python-nectar-1.5.2-1.el7.noarch
Aug 18 19:22:41 Updated: python-pulp-repoauth-2.8.6-1.el7.noarch
Aug 18 19:22:41 Updated: python-pulp-rpm-common-2.8.6-1.el7.noarch
Aug 18 19:22:42 Updated: python-pulp-oid_validation-2.8.6-1.el7.noarch
Aug 18 19:22:42 Updated: python-pulp-puppet-common-2.8.6-1.el7.noarch
Aug 18 19:22:42 Updated: python-pulp-docker-common-2.0.2-1.el7.noarch
Aug 18 19:22:42 Updated: python-pulp-agent-lib-2.8.6-1.el7.noarch
Aug 18 19:23:51 Updated: pulp-selinux-2.8.6-1.el7.noarch
Aug 18 19:23:51 Updated: python-django-bash-completion-1.8.14-1.el7.noarch
Aug 18 19:23:59 Updated: python-django-1.8.14-1.el7.noarch
Aug 18 19:24:05 Updated: pulp-server-2.8.6-1.el7.noarch
Aug 18 19:24:05 Updated: python-pulp-streamer-2.8.6-1.el7.noarch
Aug 18 19:24:05 Updated: pulp-puppet-plugins-2.8.6-1.el7.noarch
Aug 18 19:24:06 Updated: pulp-docker-plugins-2.0.2-1.el7.noarch
Aug 18 19:24:06 Updated: pulp-rpm-plugins-2.8.6-1.el7.noarch
Aug 18 19:24:06 Updated: pulp-rpm-handlers-2.8.6-1.el7.noarch
Aug 18 19:24:08 Updated: 1:v8-
Aug 18 19:24:11 Updated: python-crane-2.0.1-1.el7.noarch
Aug 18 19:24:11 Updated: 1:mariadb-libs-5.5.50-1.el7_2.x86_64

One of these broke httpd on the capsule. Message from the failed service:

AH00526: Syntax error on line 28 of /etc/httpd/conf.d/pulp.conf:
Name duplicates previous WSGI daemon definition.

Line 28 in pulp.conf is

WSGIDaemonProcess pulp user=apache group=apache processes=3 display-name=%{GROUP}

This line is duplicated in 05-pulp-https.conf

[root@wellcapsuleext conf.d]# cat 05-pulp-https.conf | grep WSGIDaemonProcess
WSGIDaemonProcess pulp user=apache group=apache processes=3 display-name=%{GROUP}

According to removing the 05-pulp-https.conf is the answer - doing that does allow httpd to start, but then in the katello GUI browsing to the capsule gives an SSL error for the content sync status. Swapping the two files so that 05-pulp-https.conf was in /etc/httpd/conf.d instead of pulp.conf still allows httpd to start, and gives no error in the katello GUI

I have attached both 05-pulp-https.conf and pulp.conf

05-pulp-https.conf 05-pulp-https.conf 2.05 KB Dylan Baars, 08/21/2016 05:56 PM
pulp.conf pulp.conf 3.58 KB Dylan Baars, 08/21/2016 05:56 PM


#1 Updated by Justin Sherrill over 5 years ago

  • Legacy Backlogs Release (now unused) changed from 143 to 184

#2 Updated by Justin Sherrill over 5 years ago

  • Status changed from New to Rejected
  • Assignee set to Justin Sherrill
  • Legacy Backlogs Release (now unused) changed from 184 to 166

I believe 'foreman-installer' just needed to be run after upgrading. This is common after doing a 'yum update' as different updates can drop files in /etc/httpd/conf.d/ which conflict with what puppet is managing.

Going ahead and closing, feel free to comment if you feel different.

Also available in: Atom PDF