|
# This file is part of Buildbot. Buildbot is free software: you can
# redistribute it and/or modify it under the terms of the GNU General Public
# License as published by the Free Software Foundation, version 2.
#
# This program is distributed in the hope that it will be useful, but WITHOUT
# ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more
# details.
#
# You should have received a copy of the GNU General Public License along with
# this program; if not, write to the Free Software Foundation, Inc., 51
# Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
#
# Copyright Buildbot Team Members
import time
import traceback
import inspect
import shutil
import os
import sqlalchemy as sa
import tempfile
from buildbot.process import metrics
from twisted.internet import reactor, threads
from twisted.python import threadpool, log
# set this to True for *very* verbose query debugging output; this can
# be monkey-patched from master.cfg, too:
# from buildbot.db import pool
# pool.debug = True
debug = False
_debug_id = 1
def timed_do_fn(f):
"""Decorate a do function to log before, after, and elapsed time,
with the name of the calling function. This is not speedy!"""
def wrap(callable, *args, **kwargs):
global _debug_id
# get a description of the function that called us
st = traceback.extract_stack(limit=2)
file, line, name, _ = st[0]
# and its locals
frame = inspect.currentframe(1)
locals = frame.f_locals
# invent a unique ID for the description
id, _debug_id = _debug_id, _debug_id+1
descr = "%s-%08x" % (name, id)
start_time = time.time()
log.msg("%s - before ('%s' line %d)" % (descr, file, line))
for name in locals:
if name in ('self', 'thd'):
continue
log.msg("%s - %s = %r" % (descr, name, locals[name]))
# wrap the callable to log the begin and end of the actual thread
# function
def callable_wrap(*args, **kargs):
log.msg("%s - thd start" % (descr,))
try:
return callable(*args, **kwargs)
finally:
log.msg("%s - thd end" % (descr,))
d = f(callable_wrap, *args, **kwargs)
def after(x):
end_time = time.time()
elapsed = (end_time - start_time) * 1000
log.msg("%s - after (%0.2f ms elapsed)" % (descr, elapsed))
return x
d.addBoth(after)
return d
wrap.__name__ = f.__name__
wrap.__doc__ = f.__doc__
return wrap
class DBThreadPool(threadpool.ThreadPool):
running = False
# Some versions of SQLite incorrectly cache metadata about which tables are
# and are not present on a per-connection basis. This cache can be flushed
# by querying the sqlite_master table. We currently assume all versions of
# SQLite have this bug, although it has only been observed in 3.4.2. A
# dynamic check for this bug would be more appropriate. This is documented
# in bug #1810.
__broken_sqlite = False
def __init__(self, engine, verbose=False):
# verbose is used by upgrade scripts, and if it is set we should print
# messages about versions and other warnings
log_msg = log.msg
if verbose:
def log_msg(m):
print m
pool_size = 5
# If the engine has an C{optimal_thread_pool_size} attribute, then the
# maxthreads of the thread pool will be set to that value. This is
# most useful for SQLite in-memory connections, where exactly one
# connection (and thus thread) should be used.
if hasattr(engine, 'optimal_thread_pool_size'):
pool_size = engine.optimal_thread_pool_size
threadpool.ThreadPool.__init__(self,
minthreads=1,
maxthreads=pool_size,
name='DBThreadPool')
self.engine = engine
if engine.dialect.name == 'sqlite':
vers = self.get_sqlite_version()
if vers < (3,7):
log_msg("Using SQLite Version %s" % (vers,))
log_msg("NOTE: this old version of SQLite does not support "
"WAL journal mode; a busy master may encounter "
"'Database is locked' errors. Consider upgrading.")
if vers < (3,4):
log_msg("NOTE: this old version of SQLite is not "
"supported.")
raise RuntimeError("unsupported SQLite version")
brkn = self.__broken_sqlite = self.detect_bug1810()
if brkn:
log_msg("Applying SQLite workaround from Buildbot bug #1810")
self._start_evt = reactor.callWhenRunning(self._start)
# patch the do methods to do verbose logging if necessary
if debug:
self.do = timed_do_fn(self.do)
self.do_with_engine = timed_do_fn(self.do_with_engine)
def _start(self):
self._start_evt = None
if not self.running:
self.start()
self._stop_evt = reactor.addSystemEventTrigger(
'during', 'shutdown', self._stop)
self.running = True
def _stop(self):
self._stop_evt = None
self.stop()
self.engine.dispose()
self.running = False
def shutdown(self):
"""Manually stop the pool. This is only necessary from tests, as the
pool will stop itself when the reactor stops under normal
circumstances."""
if not self._stop_evt:
return # pool is already stopped
reactor.removeSystemEventTrigger(self._stop_evt)
self._stop()
# Try about 170 times over the space of a day, with the last few tries
# being about an hour apart. This is designed to span a reasonable amount
# of time for repairing a broken database server, while still failing
# actual problematic queries eventually
BACKOFF_START = 1.0
BACKOFF_MULT = 1.05
MAX_OPERATIONALERROR_TIME = 3600*24 # one day
def __thd(self, with_engine, callable, args, kwargs):
# try to call callable(arg, *args, **kwargs) repeatedly until no
# OperationalErrors occur, where arg is either the engine (with_engine)
# or a connection (not with_engine)
backoff = self.BACKOFF_START
start = time.time()
while True:
if with_engine:
arg = self.engine
else:
arg = self.engine.contextual_connect()
if self.__broken_sqlite: # see bug #1810
arg.execute("select * from sqlite_master")
try:
try:
rv = callable(arg, *args, **kwargs)
assert not isinstance(rv, sa.engine.ResultProxy), \
"do not return ResultProxy objects!"
except sa.exc.OperationalError, e:
text = e.orig.args[0]
if not isinstance(text, basestring):
raise
if "Lost connection" in text \
or "database is locked" in text:
# see if we've retried too much
elapsed = time.time() - start
if elapsed > self.MAX_OPERATIONALERROR_TIME:
raise
metrics.MetricCountEvent.log(
"DBThreadPool.retry-on-OperationalError")
log.msg("automatically retrying query after "
"OperationalError (%ss sleep)" % backoff)
# sleep (remember, we're in a thread..)
time.sleep(backoff)
backoff *= self.BACKOFF_MULT
# and re-try
continue
else:
raise
finally:
if not with_engine:
arg.close()
break
return rv
def do(self, callable, *args, **kwargs):
return threads.deferToThreadPool(reactor, self,
self.__thd, False, callable, args, kwargs)
def do_with_engine(self, callable, *args, **kwargs):
return threads.deferToThreadPool(reactor, self,
self.__thd, True, callable, args, kwargs)
def detect_bug1810(self):
# detect buggy SQLite implementations; call only for a known-sqlite
# dialect
try:
import pysqlite2.dbapi2 as sqlite
sqlite = sqlite
except ImportError:
import sqlite3 as sqlite
tmpdir = tempfile.mkdtemp()
dbfile = os.path.join(tmpdir, "detect_bug1810.db")
def test(select_from_sqlite_master=False):
conn1 = None
conn2 = None
try:
conn1 = sqlite.connect(dbfile)
curs1 = conn1.cursor()
curs1.execute("PRAGMA table_info('foo')")
conn2 = sqlite.connect(dbfile)
curs2 = conn2.cursor()
curs2.execute("CREATE TABLE foo ( a integer )")
if select_from_sqlite_master:
curs1.execute("SELECT * from sqlite_master")
curs1.execute("SELECT * from foo")
finally:
if conn1:
conn1.close()
if conn2:
conn2.close()
os.unlink(dbfile)
try:
test()
except sqlite.OperationalError:
# this is the expected error indicating it's broken
shutil.rmtree(tmpdir)
return True
# but this version should not fail..
test(select_from_sqlite_master=True)
shutil.rmtree(tmpdir)
return False # not broken - no workaround required
def get_sqlite_version(self):
engine = sa.create_engine('sqlite://')
conn = engine.contextual_connect()
try:
r = conn.execute("SELECT sqlite_version()")
vers_row = r.fetchone()
r.close()
except:
return (0,)
if vers_row:
try:
return tuple(map(int, vers_row[0].split('.')))
except (TypeError, ValueError):
return (0,)
else:
return (0,)
|