Bug 1630129

FailedQA:
Environment:
openstack-tripleo-common-8.6.6-2.el7ost.noarch
openstack-tripleo-common-containers-8.6.6-2.el7ost.noarch


[heat-admin@overcloud-controller-1 ~]$ sudo docker ps|grep nova_placement
74c6787d2b65        192.168.0.1:8787/rhosp13/openstack-nova-placement-api:13.0-66          "kolla_start"            21 hours ago        Up 21 hours (unhealthy)                       nova_placement

with 7f9606ff47f03162e5d7305aaca2af5d1faecd87 backported to healthcheck/common.sh the wsgi_alias
is now automatically received from the vhost config. As a result we check again a wrong url:

http://controller-0.internalapi.localdomain:8778/placement/placement/

$ git show 7f9606ff47f03162e5d7305aaca2af5d1faecd87
commit 7f9606ff47f03162e5d7305aaca2af5d1faecd87
Author: Jose Luis Franco Arza <jfrancoa>
Date:   Mon Jul 16 09:40:11 2018 +0200

    Take WSGIScriptAlias into account in docker healthcheck.
    
    Some services running as vhost, as nova_placement make
    use of an  alias in which the wsgi service is mapped.
    Currently, the nova_placement healthcheck has hardcoded
    the alias in the curl [0] request, but if this alias
    changes or other service modifies its from / to another,
    then the check will start failing again.
    
    [0] - https://github.com/openstack/tripleo-common/blob/master/healthcheck/nova-placement#L6
    
    Change-Id: I5a325be424740e80eafd6b4abd8eb4b3111740aa
    Closes-Bug: #1781623
    (cherry picked from commit f1a1c324fb149d40c960f1eaba650f0e8c71d91e)

diff --git a/healthcheck/common.sh b/healthcheck/common.sh
index 85eb9a18..ab55e4dc 100644
--- a/healthcheck/common.sh
+++ b/healthcheck/common.sh
@@ -50,9 +50,13 @@ get_url_from_vhost () {
   server_name=$(awk '/ServerName/ {print $2}' $vhost_file)
   ssl_enabled=$(awk '/SSLEngine/ {print $2}' $vhost_file)
   bind_port=$(grep -h "<VirtualHost .*>" $vhost_file | sed 's/<VirtualHost .*:\(.*\)>/\1/')
+  wsgi_alias=$(awk '/WSGIScriptAlias/ {print $2}' $vhost_file)
   proto=http
   if [[ $ssl_enabled == "on" ]]; then
     proto=https
   fi
-  echo ${proto}://${server_name}:${bind_port}/
+  if [[ $wsgi_alias != "/" ]]; then
+    wsgi_alias="${wsgi_alias}/"
+  fi
+  echo ${proto}://${server_name}:${bind_port}${wsgi_alias}
 }


The change in rocky f1a1c324fb149d40c960f1eaba650f0e8c71d91e also modifies the  healthcheck/nova-placement script, whereas the backport to queens does not:
diff --git a/healthcheck/nova-placement b/healthcheck/nova-placement
index f8e29267..1e54c93e 100755
--- a/healthcheck/nova-placement
+++ b/healthcheck/nova-placement
@@ -3,4 +3,4 @@
 . ${HEALTHCHECK_SCRIPTS:-/usr/share/openstack-tripleo-common/healthcheck}/common.sh
 
 check_url=$(get_url_from_vhost /etc/httpd/conf.d/10-placement_wsgi.conf)
-healthcheck_curl ${check_url}/placement/
+healthcheck_curl ${check_url}

We'd need to revert the patch again which we backported in this BZ.

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, 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/RHBA-2019:0068

*** Bug 1684728 has been marked as a duplicate of this bug. ***