Clark Boylan 16a4bdce02 Don't always update gitea project descriptions
There is some correlation that running the manage-projects playbook
gives our gitea fits. The bulk of the work done here is in trying to
update the descriptions of all projects. There isn't a good way to see
if the description is already set first so we just try and ignore
errors. This creates potentially thousands of operations all at once and
could be why things are sad.

We move these operations under the always update flag which is not set
on normal runs. If we really need to converge to a good updated state we
can manually run the playbook/role with always update set.

We also don't set a limit on the number of ThreadPoolExecutor workers
which will default to 5 * NumProcs. Could be that tuning this down would
make gitea happier.

One other thought is that we may not be using request sessions properly
for connection reuse. In particular requests notes that you need to set
stream to False or read request content to return a connection back to
the pool for reuse. We might look into this for further improvements.

Change-Id: I6e6fb1eb08303e9da7e38cf493d1871364340000
2021-03-16 13:06:16 -07:00

358 lines
14 KiB
Python
Executable File

#!/usr/bin/env python3
#
# Copyright 2019 Red Hat, Inc
#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
import concurrent.futures
import datetime
import time
import requests
import urllib.parse
# urllib3 wants to warn us that we're making insecure requests - when we tell
# it we want to make insecure requests. We know, but in this case the requests
# are not insecure.
import urllib3
urllib3.disable_warnings()
from ansible.module_utils.basic import AnsibleModule
SB_REPO = 'https://storyboard.openstack.org/#!/project/{org}/{repo}'
SB_FORMAT = 'https://storyboard.openstack.org/#!/story/{{index}}'
LP_REPO = 'https://bugs.launchpad.net/{repo}'
LP_FORMAT = 'https://bugs.launchpad.net/{repo}/+bug/{{index}}'
class Gitea(object):
def __init__(self, url, password, always_update, projects):
self.url = url
self.password = password
self.always_update = always_update
self.projects = projects
self.orgs = { f['project'].split('/')[0] for f in self.projects }
self.org_projects = {}
for org in self.orgs:
p = [ f for f in self.projects
if (f['project'].split('/')[0] == org) ]
self.org_projects[org] = p
self._log = []
self.session = requests.Session()
self.failed = False
def log(self, *args):
now = datetime.datetime.utcnow().isoformat()
self._log.append(" ".join((now,) + args))
def get_log(self):
return "\n".join(self._log)
def request(self, method, endpoint, *args, **kwargs):
resp = self.session.request(
method,
urllib.parse.urljoin(self.url, endpoint),
auth=('root', self.password),
verify=False,
*args, **kwargs)
resp.raise_for_status()
return resp
def get(self, endpoint, *args, **kwargs):
return self.request('GET', endpoint, *args, **kwargs)
def get_paginated(self, endpoint, *args, **kwargs):
params = { 'limit': 50, 'page': 1 }
results = []
gitea_data = self.get(endpoint, *args, params=params, **kwargs).json()
while gitea_data:
results.extend(gitea_data)
# Gitea paginates and returns an empty list at the end of the
# listing. 50 items is the max limit.
params['page'] += 1
gitea_data = self.get(endpoint, *args, params=params, **kwargs).json()
return results
def patch(self, endpoint, *args, **kwargs):
return self.request('PATCH', endpoint, *args, **kwargs)
def post(self, endpoint, *args, **kwargs):
return self.request('POST', endpoint, *args, **kwargs)
def put(self, endpoint, *args, **kwargs):
return self.request('PUT', endpoint, *args, **kwargs)
def get_gitea_orgs(self):
orgs = self.get_paginated("/api/v1/user/orgs")
return [f['username'] for f in orgs]
def make_gitea_org(self, org):
self.post(
'/api/v1/admin/users/root/orgs',
json=dict(username=org))
self.log("Created org:", org)
def ensure_gitea_teams(self, org):
team_list = self.get_paginated('/api/v1/orgs/{org}/teams'.format(org=org))
owner_id = [f['id'] for f in team_list if f['name'] == 'Owners'][0]
org_owners = self.get_paginated(
'/api/v1/teams/{owner_id}/members'.format(owner_id=owner_id))
if 'gerrit' not in [f['username'] for f in org_owners]:
self.put('/api/v1/teams/{owner_id}/members/gerrit'.format(
owner_id=owner_id))
self.log("Added gerrit to team:", org)
def get_org_repo_list(self, org):
return [x['full_name'] for x in
self.get_paginated('/api/v1/orgs/{org}/repos'.format(org=org))]
def get_csrf_token(self):
resp = self.get('/')
return urllib.parse.unquote(self.session.cookies.get('_csrf'))
def make_gitea_project(self, project, csrf_token):
org, repo = project['project'].split('/', 1)
repo_properties = {
'auto_init': True,
'name': repo,
'description': project.get('description', '')[:255],
# Do not use this functionality until jeepyb can do similar
# for the gerrit side. Once Gerrit and Gitea can be configured
# this could be used on new repos.
# Note we default to master to avoid relying on tool defaults
# as we currently rely on Gitea, Gerrit, and Git to all be in
# sync which may not be the case going forward.
'default_branch': project.get('default-branch', 'master'),
'private': False,
'readme': 'Default',
}
resp = self.post(
'/api/v1/orgs/{org}/repos'.format(org=org),
json=repo_properties)
self.log("Created repo:", project['project'])
def update_gitea_project_settings(self, project, csrf_token):
org, repo = project['project'].split('/', 1)
if project.get('use-storyboard'):
external_tracker_url = SB_REPO.format(org=org, repo=repo)
tracker_url_format = SB_FORMAT
elif project.get('groups'):
external_tracker_url = LP_REPO.format(repo=project['groups'][0])
tracker_url_format = LP_FORMAT.format(repo=project['groups'][0])
else:
external_tracker_url = LP_REPO.format(repo=repo)
tracker_url_format = LP_FORMAT.format(repo=repo)
for count in range(0, 5):
try:
self.post(
'/{org}/{repo}/settings'.format(org=org, repo=repo),
data=dict(
_csrf=csrf_token,
action='advanced',
# enable_pulls is not provided, which disables it
# enable_wiki is not provided, which disables it
enable_external_wiki=False,
external_wiki_url='',
# enable_issues is on so that issue links work
enable_issues='on',
enable_external_tracker=True,
external_tracker_url=external_tracker_url,
tracker_url_format=tracker_url_format,
tracker_issue_style='numeric',
),
allow_redirects=False)
# Set allow_redirects to false because gitea returns
# with a 302 on success, and we don't need to follow
# that.
self.log("Updated tracker url:", external_tracker_url)
return
except requests.exceptions.HTTPError as e:
time.sleep(3)
raise Exception("Could not update tracker url")
def update_gitea_project_branches(self, project, csrf_token):
org, repo = project['project'].split('/', 1)
for count in range(0, 5):
try:
self.post(
'/{org}/{repo}/settings/branches'.format(
org=org, repo=repo),
data=dict(
_csrf=csrf_token,
action='default_branch',
branch='master',
),
allow_redirects=False)
# Set allow_redirects to false because gitea returns
# with a 302 on success, and we don't need to follow
# that.
self.log("Set master branch:", project['project'])
return
except requests.exceptions.HTTPError as e:
time.sleep(3)
raise Exception("Could not update branch settings")
def update_gitea_project_description(self, project, csrf_token):
org, repo = project['project'].split('/', 1)
description = project.get('description', '')[:255]
if description:
description_update = {
'description': description,
}
try:
resp = self.patch(
'/api/v1/repos/{org}/{repo}'.format(org=org, repo=repo),
json=description_update)
# Commented out as there is no good way to log only those
# projects which have an updated description and as a result
# this is noisy.
#self.log("Set description for:", project['project'])
except Exception as e:
# Updating descriptions is best effort as we may fail due to
# gitea bugs, but such a failure isn't critical.
self.log("Failed to set desciption for:",
project['project'], str(e))
def make_projects(self, projects, gitea_repos, csrf_token,
settings_thread_pool, branches_thread_pool, futures):
for project in projects:
create = False
if project['project'] not in gitea_repos:
try:
self.get('/' + project['project'])
except requests.HTTPError:
# If the project isn't in the listing we do an explicit
# check for its existence. This is because gitea repo
# listings require pagination and they don't use stable
# sorting and that causes problems reliably producing a
# complete repo list. If we cannot find the project
# then create it.
create = True
if create:
# TODO: use threadpool when we're running with
# https://github.com/go-gitea/gitea/pull/7493
self.make_gitea_project(project, csrf_token)
if create or self.always_update:
futures.append(settings_thread_pool.submit(
self.update_gitea_project_settings,
project, csrf_token))
futures.append(branches_thread_pool.submit(
self.update_gitea_project_branches,
project, csrf_token))
if self.always_update:
# If we are not creating, but are trying to always update
# then we update the project description.
futures.append(settings_thread_pool.submit(
self.update_gitea_project_description,
project, csrf_token))
def run(self):
futures = []
gitea_orgs = self.get_gitea_orgs()
gitea_repos = []
for org in self.orgs:
if org not in gitea_orgs:
self.make_gitea_org(org)
self.ensure_gitea_teams(org)
gitea_repos.extend(self.get_org_repo_list(org))
csrf_token = self.get_csrf_token()
# We can create repos in parallel, as long as all the repos
# for the same org are in series (due to database contention,
# until https://github.com/go-gitea/gitea/pull/7493 is
# merged). It doesn't help to have more than 2 since
# openstack is the largest and everything else combined is
# less than that.
org_thread_pool = concurrent.futures.ThreadPoolExecutor(max_workers=2)
settings_thread_pool = concurrent.futures.ThreadPoolExecutor()
branches_thread_pool = concurrent.futures.ThreadPoolExecutor()
# The very first update to the repo_unit table needs to happen
# without any other actions in parallel, otherwise a lock will
# be held for a significant amount of time causing requests to
# back up (and some to fail). Work through the project list
# in series until we find the first that updates the project
# settings (this will be the first with any significant work).
org_task_lists = []
for org, projects in self.org_projects.items():
org_task_lists.append(projects)
first_settings = False
for task_list in org_task_lists:
while task_list:
project = task_list.pop(0)
self.make_projects([project], gitea_repos, csrf_token,
settings_thread_pool, branches_thread_pool,
futures)
if len(futures) > 1:
first_settings = True
self.wait_for_futures(futures)
futures = []
if first_settings:
break
# Once that is done, we can parallelize the rest. Sort the
# org task lists by length so that we pack them into our two
# threads efficiently.
sorted_task_lists = sorted(
org_task_lists, key=lambda x: len(x), reverse=True)
for projects in sorted_task_lists:
futures.append(org_thread_pool.submit(
self.make_projects,
projects, gitea_repos, csrf_token, settings_thread_pool,
branches_thread_pool, futures))
self.wait_for_futures(futures)
def wait_for_futures(self, futures):
for f in futures:
try:
r = f.result()
except Exception as e:
self.log(str(e))
self.failed = True
def ansible_main():
module = AnsibleModule(
argument_spec=dict(
url=dict(required=True),
password=dict(required=True, no_log=True),
projects=dict(required=True, type='list'),
always_update=dict(type='bool', default=True),
)
)
p = module.params
gitea = Gitea(
url=p.get('url'),
password=p.get('password'),
always_update=p.get('always_update'),
projects=p.get('projects'),
)
try:
gitea.run()
except Exception as e:
module.fail_json(msg=str(e), changed=True)
log = gitea.get_log()
if gitea.failed:
module.fail_json(msg="Failure during repo creation, see log",
changed=bool(log), log=log)
module.exit_json(changed=bool(log), log=log)
if __name__ == '__main__':
ansible_main()