fedora-infrastructure

#7152 what are the most used ansible roles? trying to update and contribute.
Closed: Invalid Opened by jlozadad.

Closed: Invalid
Reopen Issue

looking for feedback on what ansible roles are mostly used so I can concentrate on a list. I'm going through right now on the postgresql_server role and going to make some changes.

I would say the most used roles would be the base one.. but here is how I tried to itemize it.

git clone ansible # do as needed
cd ansible/playbooks
[smooge@batcave01 playbooks (master)]$ find . -type f -print | xargs grep -l roles:  | xargs awk '/roles:/{flag=1;next}/:/{flag=0}flag; /role:/{print " - "$NF}' | awk '{for (i=1; i<=NF; i++){ if ($i~/-/ || $i~/role:/){print $(i+1)};} }' | sort | uniq -c | sort -bnr | head -n 25
    136 }
    124 
    108 httpd/website
    105 base
     99 nagios_client
     96 rkhunter
     95 hosts
     94 fas_client
     93 httpd/reverseproxy
     87 sudo
     80 collectd/base
     74 httpd/redirect
     54 openshift/object
     42 openvpn/client,
     42 httpd/redirectmatch
     40 fedmsg/base
     33 keytab/service
     31 rsyncd
     27 basessh
     23 nfs/client
     19 httpd/certificate
     17 openshift/secret-file
     17 collectd/fedmsg-service
     14 fedmsg/hub
     14 apache

The above may not be correct so please double check yourself and come up with a better pattern match.

Possibly a prettier, if more verbose, way to do this (though I truly envy @smooge's awk abilities)

import os
import yaml
paths = [
    '/srv/web/infra/ansible/playbooks/groups',
    '/srv/web/infra/ansible/playbooks/hosts'
]
counts = {}
files = []
for path in paths:
    d = os.listdir(path)
    for f in d:
        if f.endswith('.yml'):
            files.append(path + '/' + f)
for f in files:
    with open(f, 'r') as stream:
        try:
            q = yaml.load(stream)
            things_with_roles = [x for x in q if 'roles' in x]
            for thing in things_with_roles:
                for role in thing['roles']:
                    if isinstance(role, dict):
                        if role['role'] in counts:
                            counts[role['role']] += 1
                        else:
                            counts[role['role']] = 1
                    else:
                        if role in counts:
                            counts[role] += 1
                        else:
                            counts[role] = 1
        except yaml.YAMLError as exc:
            print exc
for k in reversed(sorted(counts, key=counts.get)):
    print "%s\t\t%s" % (k,counts[k])

gives

base        106
nagios_client       102
sudo        101
hosts       100
fas_client      99
rkhunter        98
collectd/base       88
openvpn/client      64
fedmsg/base     51
nfs/client      45
rsyncd      35
keytab/service      32
mod_wsgi        28
basessh     26
apache      22
collectd/fedmsg-service     18
fedmsg/hub      15
osbs-secret     9
postgresql_server       6
mariadb_server      5
push-docker     5
httpd/mod_ssl       5
cgit/make_pkgs_list     4
httpd/certificate       4
builder_repo        4

I'm going to close this since there is no actionable item here.
Edited by codeblock

Metadata Update from @codeblock:
- Issue close_status updated to: Invalid
- Issue status updated to: Closed (was: Open)

Caveat: This doesn't actually take number of hosts into account, for the group playbooks. So there might be only one playbook that calls in the foo role, but that playbook might control 50 hosts. My script will only count that as one.

In other words, this is a count of playbooks that use each role, not hosts that have that role applied.
Metadata
None
None
None
None
Lowest
Static archive. pagure.io was sunset in 2026; this is a read-only snapshot.
Red Hat Logo

Fedora is sponsored by Red Hat. Learn more about the relationship between Red Hat and Fedora »

© Red Hat, Inc. and others.