On production, my Flask / SQLAlchemy app randomly throws psycopg2.OperationalError: server closed the connection unexpectedly on RDS Aurora. Until I track down the issue, I would like my unit tests to ensure that I handle this properly, e.g. that my rollback mechanism is effective, etc.
Right now, I mock the Session.commit method to throw a fake exception, but this does not (afaik?) leave the session in a failed state that needs actual rollback.
What are some reliable ways to simulate an actual connection failure in my local, Docker Compose-based development environment?
Rollback with psycopg2
conn = psycopg2.connect(
f"host='{credentials['host']}' port={credentials['port']} dbname='{credentials['dbname']}' user='{credentials['user']}' password='{credentials['password']}'"
)
print("Connected to database")
print("Setting autocommit false")
conn.autocommit = False
func_name= 'procedure_update_users'
print(f"Executing function {func_name}...")
try:
cur = conn.cursor()
query = f"""SELECT {func_name}();"""
print(query)
cur.execute(query)
row = cur.fetchone()
conn.commit()
print(f"Function {func_name} executed.")
except Exception as e:
print(e)
conn.rollback()
raise Exception(e)
To get OperationalError I did:
try:
cur.execute("LOCK TABLE mytable IN ACCESS EXCLUSIVE MODE NOWAIT")
except psycopg2.OperationalError as e:
cur.execute("rollback")
raise Exception(e)
psycopg2.OperationalError: fe_sendauth: no password supplied
I am currently trying to build an application that manages multiple databases. Since the app will be managing data in 30+ databases I am attempting to generate DATABASE_ROUTERS in the settings file. I cannot directly import the db model into the settings file. I get this error:
django.core.exceptions.AppRegistryNotReady: Apps aren't loaded yet.
This error makes since. Is there a way I can control the sequence of events so that I have access to the database before all of the settings are established on execution? My goal is to automate database connections pulling relevant data from a DB and generate the DATABASE_ROUTERS and DATABASES within the setting file. Is this even possible? Is there a package that I can download that does exactly this?
If you do not know what I am asking please do not down vote just ask me to elaborate.
I was able to figure out how to query the data I needed from my database and import it into the settings file. I created the script below. Keep in mind this can be improved, this is just something I modified from here. This directly queries data from my test db (sqlite3). I use postgreSQL in production. This script should work, with some modification, with PostgreSQL.
As you can see below I am storing the data in dictionaries that is then stored in a list. I then import that list of dictionaries into my settings file. From there I can loop through the list and create my DATABASE_ROUTERS and DATABASES dynamically from the database. I was also able to generate router Classes in my routers.py file by importing the same list. Please comment below if you need me to elaborate further.
import sqlite3
from sqlite3 import Error
dbs = []
def create_connection(db_file):
""" create a database connection to the SQLite database
specified by the db_file
:param db_file: database file
:return: Connection object or None
"""
try:
conn = sqlite3.connect(db_file)
return conn
except Error as e:
print(e)
return None
def select_all_data(conn):
"""
Query all rows in the table
:param conn: the Connection object
:return:
"""
cur = conn.cursor()
cur.execute("SELECT * FROM fund_table")
rows = cur.fetchall()
for row in rows:
print(row)
def select_name_and_db(conn):
"""
Query table by fund_name and db_name
:param conn: the Connection object
:return:
"""
cur = conn.cursor()
cur.execute("SELECT fund_name, db_name FROM fund_table")
rows = cur.fetchall()
for row in rows:
dbs.append({"fund_name": row[0], "db_name": row[1]})
return dbs
def main():
database = "edb.sqlite3"
# create a database connection
conn = create_connection(database)
with conn:
""" select_all_data(conn) """
select_name_and_db(conn)
main()
Make one function that loads this variables and make it async, so after you app is ready you load it, but im not sure is this will work properly
https://hackernoon.com/asynchronous-python-45df84b82434
Dirty Sollution is make 1 file for each BD and you call your settings based in what BD gonna work...
I am using the python module called PyAthenaJDBC in order to query Athena using the provided JDBC driver.
Here is the link : https://pypi.python.org/pypi/PyAthenaJDBC/
I have been facing some persistent issue. I keep getting this java error whenever I use the Athena connection twice in a row.
As a matter of fact, I was able to connect to Athena, show databases, create new tables and even query the content. I am building an application using Django and running its server to use Athena
However, I am obliged to re-run the server in order for the Athena connection to work once again,
Here is a glimpse of the class I have built
import os
import configparser
import pyathenajdbc
#Get aws credentials for the moment
aws_config_file = '~/.aws/config'
Config = configparser.ConfigParser()
Config.read(os.path.expanduser(aws_config_file))
access_key_id = Config['default']['aws_access_key_id']
secret_key_id = Config['default']['aws_secret_access_key']
BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
athena_jdbc_driver_path = BASE_DIR + "/lib/static/AthenaJDBC.jar"
log_path = BASE_DIR + "/lib/static/queries.log"
class PyAthenaLoader():
def __init__(self):
pyathenajdbc.ATHENA_JAR = athena_jdbc_driver_path
def connecti(self):
self.conn = pyathenajdbc.connect(
s3_staging_dir="s3://aws-athena-query-results--us-west-2",
access_key=access_key_id,
secret_key=secret_key_id,
#profile_name = "default",
#credential_file = aws_config_file,
region_name="us-west-2",
log_path=log_path,
driver_path=athena_jdbc_driver_path
)
def databases(self):
dbs = self.query("show databases;")
return dbs
def tables(self, database):
tables = self.query("show tables in {0};".format(database))
return tables
def create(self):
self.connecti()
try:
with self.conn.cursor() as cursor:
cursor.execute(
"""CREATE EXTERNAL TABLE IF NOT EXISTS sales4 (
Day_ID date,
Product_Id string,
Store_Id string,
Sales_Units int,
Sales_Cost float,
Currency string
)
ROW FORMAT SERDE 'org.apache.hadoop.hive.serde2.lazy.LazySimpleSerDe'
WITH SERDEPROPERTIES (
'serialization.format' = '|',
'field.delim' = '|',
'collection.delimm' = 'undefined',
'mapkey.delim' = 'undefined'
) LOCATION 's3://athena-internship/';
""")
res = cursor.description
finally:
self.conn.close()
return res
def query(self, req):
self.connecti()
try:
with self.conn.cursor() as cursor:
cursor.execute(req)
print(cursor.description)
res = cursor.fetchall()
finally:
self.conn.close()
return res
def info(self):
res = []
for i in dir(pyathenajdbc):
temp = i + ' = ' + str(dic[i])
#print(temp)
res.append(temp)
return res
Example of usage :
def test(request):
athena = jdbc.PyAthenaLoader()
res = athena.query('Select * from sales;')
return render(request, 'test.html', {'data': res})
Works just fine!
However refreshing the page would cause this error :
Error
Note that I am using a local .jar file: I thought that would solve the issue but I was wrong
Even if I remove the path of the JDBC driver and let the module download it from s3, the error persists:
File "/home/tewfikghariani/.virtualenvs/venv/lib/python3.4/site-packages/pyathenajdbc/connection.py", line 69, in init
ATHENA_CONNECTION_STRING.format(region=self.region_name, schema=schema_name), props)
jpype._jexception.java.sql.SQLExceptionPyRaisable:
java.sql.SQLException: No suitable driver found for
jdbc:awsathena://athena.us-west-2.amazonaws.com:443/hive/default/
Furthermore, when I run the module on its own, it works just fine.
When I set multiple connection inside my view before rendering the template, that works just fine as well.
I guess the issue is related to the django view, once one of the views is performing a connection with athena, the next connection is not possible anymore and the error is raised unless I restart the server
Any help? If other details are missing I will provide them immediately.
Update:
After posting the issue in github, the author solved this problem and released a new version that works perfectly.
It was a multi-threading problem with JPype.
Question answered!
ref : https://github.com/laughingman7743/PyAthenaJDBC/pull/8
am new in python and i made a program to add data from a form to database am using wamp server here is my app.py code
import web
import MySQLdb
urls = (
'/', 'Index'
)
app = web.application(urls, globals())
render = web.template.render('templates/', base = "layout")
class Index(object):
def GET(self):
return render.hello_form()
def POST(self): #Open database connection
try:
db = MySQLdb.connect("localhost", "root", "", "testdb")
# prepare a cursor object using cursor() method
cursor = db.cursor()
# Prepare SQL query to INSERT a record into the database.
sql = "INSERT INTO details(name, address,) VALUES ("[self.name.text, self.address.text])""
try:
# Execute the SQL command
cursor.execute(sql)
# Commit your changes in the database
db.commit()
except:
# Rollback in case there is any error
db.rollback()
# disconnect from server
db.close()
finally:
form = web.input(name = "Nobody", greet = "Hello")
greeting = "%s, %s" % (form.name, form.address)
return render.index(greeting = greeting)
if __name__ == '__main__':
main()
and my error is Expected an indended block at line 12(return.render.hello_form)
can any one help pls..?
Scope is determined in python by indentation. As a result, you need to indent the try statement as it is inside the function POST. Also, the formatting on finally (and db.close()) looks wrong. Check out the this SO question, this guide and the Pep 8 guide on best practice code formatting. Also, be careful not to mix spaces and tabs as this can cause problems.
I'm trying to run
./manage.py test
But it tells me
Got an error creating the test database: permission denied to create database
Obviously it doesn't have permission to create the database, but I'm on a shared server, so there's not much I can do about that. I can create a new database through the control panel but I don't think there's any way I can let Django do it automatically.
So, can't I create the test database manually and instead tell Django to flush it every time, rather than recreating the whole thing?
I had a similar issue. But I wanted Django to just bypass the creation of a test database for one of my instances (it is not a mirror tough). Following Mark's suggestion, I created a custom test runner, as follows
from django.test.simple import DjangoTestSuiteRunner
class ByPassableDBDjangoTestSuiteRunner(DjangoTestSuiteRunner):
def setup_databases(self, **kwargs):
from django.db import connections
old_names = []
mirrors = []
for alias in connections:
connection = connections[alias]
# If the database is a test mirror, redirect its connection
# instead of creating a test database.
if connection.settings_dict['TEST_MIRROR']:
mirrors.append((alias, connection))
mirror_alias = connection.settings_dict['TEST_MIRROR']
connections._connections[alias] = connections[mirror_alias]
elif connection.settings_dict.get('BYPASS_CREATION','no') == 'no':
old_names.append((connection, connection.settings_dict['NAME']))
connection.creation.create_test_db(self.verbosity, autoclobber=not self.interactive)
return old_names, mirrors
Then I created an extra dict entry in one of my databases entries inside settings.py, 'BYPASS_CREATION':'yes',
Finally, I configured a new TestRunner with
TEST_RUNNER = 'auth.data.runner.ByPassableDBDjangoTestSuiteRunner'
I would suggest using sqlite3 for testing purposes while keeping on using mysql/postgres/etc for production.
This can be achieved by placing this in your settings file:
if 'test' in sys.argv:
DATABASES['default'] = {'ENGINE': 'django.db.backends.sqlite3'}
see Running django tests with sqlite
a temporary sqlite database file will be created in your django project home which you will have write access to. The other advantage is that sqlite3 is much faster for testing. You may however run in to problems if you are using any mysql/postgres specific raw sql (which you should try to avoid anyway).
I think a better solution might be to define your own test runner.
I added this to the comments above but it got kind of lost - recent changes to webfaction make this MUCH easier. You can now create new private database instances.
Follow the instructions there, and when creating a new user make sure to give them the permission to ALTER USER new_username CREATEDB;.
You probably also should change the default cron settings so they don't try to check if this database is up and runnings as frequently.
You could use django-nose as your TEST_RUNNER. Once installed, if you pass the following environment variable, it will not delete and re-create the database (create it manually yourself first).
REUSE_DB=1 ./manage.py test
You can also add the following to settings.py so you don't have to write REUSE_DB=1 every time you want to run tests:
os.environ['REUSE_DB'] = "1"
Note: this will also leave all your tables in the databases which means test setup will be a little quicker, but you will have to manually update the tables (or delete and re-create the database yourself) when you change your models.
my variant to reusing database:
from django.test.simple import DjangoTestSuiteRunner
from django.core.management import call_command
class TestRunner(DjangoTestSuiteRunner):
def setup_databases(self, **kwargs):
from django.db import connections
settings = connections['default'].settings_dict
settings['NAME'] = settings['TEST_NAME']
settings['USER'] = settings['TEST_USER']
settings['PASSWORD'] = settings['TEST_PASSWD']
call_command('syncdb', verbosity=1, interactive=False, load_initial_data=False)
def teardown_databases(self, old_config, **kwargs):
from django.db import connection
cursor = connection.cursor()
cursor.execute('show tables;')
parts = ('DROP TABLE IF EXISTS %s;' % table for (table,) in cursor.fetchall())
sql = 'SET FOREIGN_KEY_CHECKS = 0;\n' + '\n'.join(parts) + 'SET FOREIGN_KEY_CHECKS = 1;\n'
connection.cursor().execute(sql)
The following is a django test suite runner to create database using Webfaction XML-RPC API. Note, setting up the database using the API may take up to a minute, and the script may appear to be stuck momentarily, just wait for a little while.
NOTE: there is a security risk of having control panel password in the webfaction server, because someone breaching into your web server SSH could take over your Webfaction account. If that is a concern, set USE_SESSKEY to True and use the fabric script below this script to pass a session id to the server. The session key expires in 1 hour from the last API call.
File test_runner.py: in the server, you need to configure ./manage.py test to use WebfactionTestRunner
"""
This test runner uses Webfaction XML-RPC API to create and destroy database
"""
# you can put your control panel username and password here.
# NOTE: there is a security risk of having control panel password in
# the webfaction server, because someone breaching into your web server
# SSH could take over your Webfaction account. If that is a concern,
# set USE_SESSKEY to True and use the fabric script below this script to
# generate a session.
USE_SESSKEY = True
# CP_USERNAME = 'webfactionusername' # required if and only if USE_SESSKEY is False
# CP_PASSWORD = 'webfactionpassword' # required if and only if USE_SESSKEY is False
import sys
import os
from django.test.simple import DjangoTestSuiteRunner
from django import db
from webfaction import Webfaction
def get_sesskey():
f = os.path.expanduser("~/sesskey")
sesskey = open(f).read().strip()
os.remove(f)
return sesskey
if USE_SESSKEY:
wf = Webfaction(get_sesskey())
else:
wf = Webfaction()
wf.login(CP_USERNAME, CP_PASSWORD)
def get_db_user_and_type(connection):
db_types = {
'django.db.backends.postgresql_psycopg2': 'postgresql',
'django.db.backends.mysql': 'mysql',
}
return (
connection.settings_dict['USER'],
db_types[connection.settings_dict['ENGINE']],
)
def _create_test_db(self, verbosity, autoclobber):
"""
Internal implementation - creates the test db tables.
"""
test_database_name = self._get_test_db_name()
db_user, db_type = get_db_user_and_type(self.connection)
try:
wf.create_db(db_user, test_database_name, db_type)
except Exception as e:
sys.stderr.write(
"Got an error creating the test database: %s\n" % e)
if not autoclobber:
confirm = raw_input(
"Type 'yes' if you would like to try deleting the test "
"database '%s', or 'no' to cancel: " % test_database_name)
if autoclobber or confirm == 'yes':
try:
if verbosity >= 1:
print("Destroying old test database '%s'..."
% self.connection.alias)
wf.delete_db(test_database_name, db_type)
wf.create_db(db_user, test_database_name, db_type)
except Exception as e:
sys.stderr.write(
"Got an error recreating the test database: %s\n" % e)
sys.exit(2)
else:
print("Tests cancelled.")
sys.exit(1)
db.close_connection()
return test_database_name
def _destroy_test_db(self, test_database_name, verbosity):
"""
Internal implementation - remove the test db tables.
"""
db_user, db_type = get_db_user_and_type(self.connection)
wf.delete_db(test_database_name, db_type)
self.connection.close()
class WebfactionTestRunner(DjangoTestSuiteRunner):
def __init__(self, *args, **kwargs):
# Monkey patch BaseDatabaseCreation with our own version
from django.db.backends.creation import BaseDatabaseCreation
BaseDatabaseCreation._create_test_db = _create_test_db
BaseDatabaseCreation._destroy_test_db = _destroy_test_db
return super(WebfactionTestRunner, self).__init__(*args, **kwargs)
File webfaction.py: this is a thin wrapper for Webfaction API, it need to be importable by both test_runner.py (in the remote server) and the fabfile.py (in the local machine)
import xmlrpclib
class Webfaction(object):
def __init__(self, sesskey=None):
self.connection = xmlrpclib.ServerProxy("https://api.webfaction.com/")
self.sesskey = sesskey
def login(self, username, password):
self.sesskey, _ = self.connection.login(username, password)
def create_db(self, db_user, db_name, db_type):
""" Create a database owned by db_user """
self.connection.create_db(self.sesskey, db_name, db_type, 'unused')
# deletes the default user created by Webfaction API
self.connection.make_user_owner_of_db(self.sesskey, db_user, db_name, db_type)
self.connection.delete_db_user(self.sesskey, db_name, db_type)
def delete_db(self, db_name, db_type):
try:
self.connection.delete_db_user(self.sesskey, db_name, db_type)
except xmlrpclib.Fault as e:
print 'ignored error:', e
try:
self.connection.delete_db(self.sesskey, db_name, db_type)
except xmlrpclib.Fault as e:
print 'ignored error:', e
File fabfile.py: A sample fabric script to generate session key, needed only if USE_SESSKEY=True
from fabric.api import *
from fabric.operations import run, put
from webfaction import Webfaction
import io
env.hosts = ["webfactionusername#webfactionusername.webfactional.com"]
env.password = "webfactionpassword"
def run_test():
wf = Webfaction()
wf.login(env.hosts[0].split('#')[0], env.password)
sesskey_file = '~/sesskey'
sesskey = wf.sesskey
try:
put(io.StringIO(unicode(sesskey)), sesskey_file, mode='0600')
# put your test code here
# e.g. run('DJANGO_SETTINGS_MODULE=settings /path/to/virtualenv/python /path/to/manage.py test --testrunner=test_runner.WebfactionTestRunner')
raise Exception('write your test here')
finally:
run("rm -f %s" % sesskey_file)
The accepted answer didn't work for me. It's so outdated, that it didn't run on my legacy codebase with djano 1.5.
I wrote a blogpost entirely describing how I solved this issue by creating an alternative test runner and changing django settings to provide all the required config and to use new test runner.
You need to specify a sqlite ENGINE when using unit tests. Open the settings.py and add the just after DATABASES section:
import sys
if 'test' in sys.argv or 'test_coverage' in sys.argv: #Covers regular testing and django-coverage
DATABASES['default']['ENGINE'] = 'django.db.backends.sqlite3'
DATABASES['default']['NAME'] = ':memory:'
Modify the following methods in django/db/backends/creation.py:
def _destroy_test_db(self, test_database_name, verbosity):
"Internal implementation - remove the test db tables."
# Remove the test database to clean up after
# ourselves. Connect to the previous database (not the test database)
# to do so, because it's not allowed to delete a database while being
# connected to it.
self._set_test_dict()
cursor = self.connection.cursor()
self.set_autocommit()
time.sleep(1) # To avoid "database is being accessed by other users" errors.
cursor.execute("""SELECT table_name FROM information_schema.tables WHERE table_schema='public'""")
rows = cursor.fetchall()
for row in rows:
try:
print "Dropping table '%s'" % row[0]
cursor.execute('drop table %s cascade ' % row[0])
except:
print "Couldn't drop '%s'" % row[0]
#cursor.execute("DROP DATABASE %s" % self.connection.ops.quote_name(test_database_name))
self.connection.close()
def _create_test_db(self, verbosity, autoclobber):
"Internal implementation - creates the test db tables."
suffix = self.sql_table_creation_suffix()
if self.connection.settings_dict['TEST_NAME']:
test_database_name = self.connection.settings_dict['TEST_NAME']
else:
test_database_name = TEST_DATABASE_PREFIX + self.connection.settings_dict['NAME']
qn = self.connection.ops.quote_name
# Create the test database and connect to it. We need to autocommit
# if the database supports it because PostgreSQL doesn't allow
# CREATE/DROP DATABASE statements within transactions.
self._set_test_dict()
cursor = self.connection.cursor()
self.set_autocommit()
return test_database_name
def _set_test_dict(self):
if "TEST_NAME" in self.connection.settings_dict:
self.connection.settings_dict["NAME"] = self.connection.settings_dict["TEST_NAME"]
if "TEST_USER" in self.connection.settings_dict:
self.connection.settings_dict['USER'] = self.connection.settings_dict["TEST_USER"]
if "TEST_PASSWORD" in self.connection.settings_dict:
self.connection.settings_dict['PASSWORD'] = self.connection.settings_dict["TEST_PASSWORD"]
Seems to work... just add the extra settings to your settings.py if you need 'em.
Simple workaround: change TEST_DATABASE_PREFIX in django/db/backends/base/creation.py as you like.