Python: MySQL: Handling timeouts

Question

I am using Python and mySQL, and there is a long lag between queries. As a result, I get an 'MySQL connection has gone away' error, that is wait_timeout is exceeded.

This has been discussed e.g. in Gracefully handling "MySQL has gone away"

but this does not specifically answer my query.

So my approach to handling this - I have wrapped all my sql execute statements in a method as -

  def __execute_sql(self,sql,cursor):
    try:
        cursor.execute(sql)

    except MySQLdb.OperationalError, e:            
        if e[0] == 2006:
            self.logger.do_logging('info','DB', "%s : Restarting db" %(e))
            self.start_database()

I have several places in the code which calls this query. The thing is, I also have several cursors, so the method invocations look like-

self.__execute_sql(sql,self.cursor_a)
self.__execute_sql(sql,self.cursor_b)

and so on

I need a way to gracefully re-execute the query after the db has been started. I could wrap the calls in an if statement, and re-execute so it would be

def __execute_sql(self,sql,cursor):
    try:
        cursor.execute(sql)
        return 1
except MySQLdb.OperationalError, e:            
    if e[0] == 2006:
        self.logger.do_logging('info','DB', "%s : Restarting db" %(e))
        self.start_database()
        return 0

and then

if (self.__execute_sql(sql,self.cursor_a) == 0):
   self.__execute_sql(sql,self.cursor_a)

But this is clunky. Is there a better way to do this? Thanks!!!

kiminoa · Accepted Answer · 2015-08-06 18:02:07Z

4

I tried Crasched's approach, which got me to a new OperationalError:

OperationalError: (2013, 'Lost connection to MySQL server during query')

My final solution was to first try the ping, and if another OperationalError was raised, to reconnect and recreate the cursor with the new connection, like so:

try:
    self.connection.ping(True)
except MySQLdb.OperationalError:
    self.connection = MySQLdb.connect(
        self.db_host,
        self.db_user,
        self.db_passwd,
        self.db_dbase,
        self.db_port)
    # reconnect your cursor as you did in __init__ or wherever    
    self.cursor = self.connection(
        MySQLdb.cursors.DictCursor)

Back in business!

Python 2.7, MySQL 5.5.41

answered Aug 6, 2015 at 18:02

kiminoa

2,7193 gold badges19 silver badges19 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Jeff Widman Over a year ago

Pinging before running a query is considered an anti-pattern that wastes resources and is unreliable: percona.com/blog/2010/05/05/…

Ben Wheeler Over a year ago

@kiminoa - Try executing the query, and catching any MySQL errors, in which case you re-try the query.

Crasched · Accepted Answer · 2013-03-08 14:07:13Z

2

I had the same problem and wanted to wrap the exception to capture it but instead I solved it by using the following. Before calling the execute, call
self.con.ping(TRUE)

http://www.neotitans.com/resources/python/mysql-python-connection-error-2006.html http://mysql-python.sourceforge.net/MySQLdb.html

I can no longer find the original source material I found this out from, but this solved the problem immediately.

answered Mar 8, 2013 at 14:07

Crasched

192 bronze badges

1 Comment

Jeff Widman Over a year ago

Pinging before running a query is considered an anti-pattern that wastes resources and is unreliable: percona.com/blog/2010/05/05/…

Ben Wheeler · Accepted Answer · 2017-08-29 03:46:32Z

I was running into mystifying "MySQL server has gone away" errors, and here is my solution.

This solution will let you retry through MySQL errors, handle pretty much any type of query, include query variables in the query str or in a separate tuple, and collect and return all of the success and error messages you encounter along the way:

def execute_query(query_str, values=None):
  # defaults
  num_affected_rows = 0
  result_rows = None
  success = False
  message = "Error executing query: {}".format(query_str)
  # run the query
  try:
    mysql_conn = get_existing_mysql_connection()
    cur = mysql_conn.cursor()
    if values == None or len(values) < 1:
      num_affected_rows = cur.execute(query_str)
    else:
      num_affected_rows = cur.execute(query_str, values)
    result_rows = cur.fetchall() # only relevant to select, but safe to run with others
    cur.close()
    mysql_conn.commit()
    success = True
    message = "Mysql success for query: {}".format(query_str)
  except BaseException as e:
    message = "Mysql error: {}; for query: {}".format(repr(e), query_str)
  return (success, num_affected_rows, result_rows, message)


def execute_query_with_retry(query_str, values=None, num_tries=3, message=""):
  # defaults
  success = False
  num_affected_rows = 0
  result_rows = None
  this_message = "Error executing query: {}".format(query_str)
  # should we still try?
  if num_tries < 1:
    this_message = "Ran out of tries for query: {}".format(query_str)
    return (False, 0, None, message + '; ' + this_message)
  num_tries_after_this = num_tries - 1
  # try to execute query
  try:
    (success, num_affected_rows, result_rows, this_message) = execute_query(query_str, values)
  except BaseException as e:
    success = False
  # handle success or failure
  if success == True:
    return (True, num_affected_rows, result_rows, message + '; ' + this_message)
  else:
    open_new_mysql_connection() # reconnect using password etc.
    return(execute_query_with_retry(query_str, values=values, num_tries=num_tries_after_this, message=(message + '; ' + this_message)))

Collectives™ on Stack Overflow

Python: MySQL: Handling timeouts

3 Answers 3

2 Comments

1 Comment

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related