add project publik-dump

This commit is contained in:
Christophe Siraut 2020-07-16 14:29:02 +02:00
parent a49f5b7d77
commit faba1dbabc
4 changed files with 209 additions and 0 deletions

1
publik-dump/.gitignore vendored Normal file
View File

@ -0,0 +1 @@
output

View File

@ -0,0 +1,19 @@
#!/usr/bin/python2
import argparse
import pickle
parser = argparse.ArgumentParser()
parser.add_argument("file")
parser.add_argument("--host")
parser.add_argument("--password")
args = parser.parse_args()
with open(args.file) as fh:
data = pickle.loads(fh.read())
data["postgresql"]["host"] = args.host
data["postgresql"]["password"] = args.password
print(data["postgresql"])
with open(args.file, "w") as fh:
fh.write(pickle.dumps(data))

View File

@ -0,0 +1,25 @@
#!/usr/bin/python3
import json
from urllib.parse import urlparse, urlsplit
from hobo.environment.models import AVAILABLE_SERVICES
from hobo.multitenant.middleware import TenantMiddleware
from django.db import connection
tenant = connection.get_tenant()
services = [{"name": "hobo", "url": tenant.domain_url, "schema": tenant.schema_name}]
for service in AVAILABLE_SERVICES:
for site in service.objects.all():
service = {
"name": site.Extra.service_id,
"url": urlsplit(site.base_url).netloc.split(":")[0],
}
if site.secondary:
service["secondary"] = True
schema = TenantMiddleware.hostname2schema(service["url"])
if service["name"] == "wcs":
schema = "wcs_%s" % schema
service["schema"] = schema
services.append(service)
print(json.dumps({"name": tenant.domain_url, "services": services}))

164
publik-dump/bin/publik_dump.py Executable file
View File

@ -0,0 +1,164 @@
#!/usr/bin/python3
import argparse
import json
import os
import pickle
import subprocess
parser = argparse.ArgumentParser()
parser.add_argument("action", default="dump", choices=["tenantinfo", "tenanturls", "dump", "restore"])
parser.add_argument("host", help="origin host")
parser.add_argument("tenant", help="hobo tenant url")
parser.add_argument("--update", action="store_true")
parser.add_argument("--target", help="destination host")
parser.add_argument("--dbtarget", help="destination host")
args = parser.parse_args()
host_folder = "output/%s" % args.host
def run(cmd):
print("+ %s" % cmd)
return subprocess.run(cmd, shell=True, check=True, stdout=subprocess.PIPE)
def get_dump_folder(service):
dump_folder = "%s/%s" % (host_folder, service["url"])
if not os.path.isdir(dump_folder):
os.mkdir(dump_folder)
return dump_folder
def dump_tenant_files(tenant):
for service in tenant["services"]:
dump_folder = get_dump_folder(service)
output = "%s/%s.tar.xz" % (dump_folder, service["url"])
run(
"ssh %s.%s 'sudo tar -C %s -Jcf - %s' > %s"
% (service["name"], args.host, service["path"], service["url"], output)
)
def restore_tenant_files(tenant):
assert run("ssh %s hostname -f" % args.target).stdout.decode().strip() == args.target
for service in tenant["services"]:
dump_folder = get_dump_folder(service)
input_file = "%s/%s.tar.xz" % (dump_folder, service["url"])
run(
"cat %s | ssh %s.%s 'sudo tar -C %s -Jxf -'"
% (input_file, service["name"], args.target, service["path"])
)
def dump_tenant_databases(tenant):
for service in tenant["services"]:
dump_folder = get_dump_folder(service)
if service["name"] == "wcs":
dump_file = "%s/%s.sql.gz" % (dump_folder, service["database"])
run(
"ssh database.%s 'sudo -u postgres pg_dump -Fc %s' > %s"
% (args.host, service["database"], dump_file)
)
else:
dump_file = "%s/%s.sql.gz" % (dump_folder, service["schema"])
run(
"ssh database.%s 'sudo -u postgres pg_dump -n %s -Fc %s' > %s"
% (args.host, service["schema"], service["database"], dump_file)
)
def restore_tenant_databases(tenant):
assert run("ssh %s hostname -f" % args.dbtarget).stdout.decode().strip() == args.dbtarget
for service in tenant["services"]:
dump_folder = get_dump_folder(service)
if service["name"] == "wcs":
dump_file = "%s/%s.sql.gz" % (dump_folder, service["database"])
run(
"ssh %s sudo -u postgres dropdb --if-exists %s"
% (args.dbtarget, service["database"])
)
run(
"""ssh %s 'sudo -u postgres createdb %s --owner wcs --template="template0" --lc-collate=fr_FR.utf8 --lc-ctype=fr_FR.utf8'"""
% (args.dbtarget, service["database"])
)
run(
"cat %s | ssh %s sudo -u postgres pg_restore -d %s"
% (dump_file, args.dbtarget, service["database"])
)
else:
dump_file = "%s/%s.sql.gz" % (dump_folder, service["schema"])
run(
"""ssh %s 'sudo -u postgres psql -c "drop schema if exists %s cascade" %s'"""
% (args.dbtarget, service["schema"], service["database"])
)
run(
"cat %s | ssh %s sudo -u postgres pg_restore -d %s"
% (dump_file, args.dbtarget, service["database"])
)
def parse_service(service):
if service["name"] == "authentic":
path = "/var/lib/authentic2-multitenant/tenants"
database = "authentic2_multitenant"
elif service["name"] == "wcs":
path = "/var/lib/wcs"
try:
wcs_config = run(
"ssh wcs.%s 'cat /var/lib/wcs/%s/config.pck'"
% (args.host, service["url"])
)
config = pickle.loads(wcs_config.stdout)
database = config["postgresql"]["database"]
except (KeyError, subprocess.CalledProcessError):
database = None
else:
database = service["name"]
path = "/var/lib/%s/tenants" % service["name"]
return dict(service, database=database, path=path)
def get_host_info():
if os.path.isfile("%s/data" % host_folder) and not args.update:
return
if not os.path.isdir(host_folder):
os.makedirs(host_folder)
run("scp bin/list_tenants.py hobo.%s:" % args.host)
output = run(
"ssh hobo.%s 'sudo -u hobo HOME=$HOME hobo-manage tenant_command "
"runscript ~/list_tenants.py --all-tenants'" % args.host
)
tenants = []
for line in output.stdout.decode().split("\n"):
if line:
tenant = json.loads(line)
for i, service in enumerate(tenant["services"]):
tenant["services"][i] = parse_service(service)
tenants.append(tenant)
with open("%s/data" % host_folder, "w") as fh:
fh.write(json.dumps(tenants, indent=4))
def get_tenant_info(tenant_name):
with open("%s/data" % host_folder) as fh:
hobos = json.loads(fh.read())
tenant_infos = [x for x in hobos if x["name"] == tenant_name]
if not len(tenant_infos) == 1:
raise(Exception('tenant not found'))
return tenant_infos[0]
if __name__ == "__main__":
if args.action == "tenantinfo":
get_host_info()
print(json.dumps(get_tenant_info(args.tenant), indent=4))
if args.action == "tenanturls":
print(' '.join([x['url'] for x in get_tenant_info(args.tenant)['services']]))
elif args.action == "dump":
get_host_info()
tenant = get_tenant_info(args.tenant)
dump_tenant_databases(tenant)
dump_tenant_files(tenant)
elif args.action == "restore":
tenant = get_tenant_info(args.tenant)
restore_tenant_files(tenant)
restore_tenant_databases(tenant)