我在 Apache 中使用 Djang 时遇到了问题。它以守护进程模式运行并使用 worker MPM。该应用程序主要是一个 restful API，用于为来自移动设备和外部 Web 服务的连接提供服务。
当服务器进程达到其最大请求数时，它会重新启动。似乎当进程重新启动时，python 进程可能会锁定。我已抓取进程的堆栈跟踪，看起来它们在导入数据库引擎时锁定了。
工作者设置：
<IfModule worker.c>
StartServers 5
MaxClients 150
MinSpareThreads 20
MaxSpareThreads 100
ThreadsPerChild 10
MaxRequestsPerChild 0
</IfModule>
和虚拟主机：
NameVirtualHost *:443
<VirtualHost *:443>
Header set Access-Control-Allow-Origin "*"
ServerName example.com
WSGIScriptAlias / /example/API/apache/django.wsgi
ErrorLog /example/API/apache/log
<Directory /example/API>
<Files django.wsgi>
Order allow,deny
Allow from all
</Files>
</Directory>
</VirtualHost>
我已经使用此处描述的方法获取了几个堆栈跟踪https://code.google.com/p/modwsgi/wiki/DebuggingTechniques#Extracting_Python_Stack_Traces。
该进程有 10 个线程。其中 9 个已经产生了这个。
# ProcessId: 10353
# ThreadID: 140479095494400
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/api/web_transaction.py", line 853, in __call__
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/api/function_trace.py", line 108, in literal_wr\
apper
File: "/usr/lib/python2.6/site-packages/django/core/handlers/wsgi.py", line 206, in __call__
File: "/usr/lib/python2.6/site-packages/django/core/handlers/base.py", line 112, in get_response
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/hooks/framework_django.py", line 492, in wrappe\
r
File: "/usr/lib/python2.6/site-packages/django/views/generic/base.py", line 69, in view
File: "/usr/lib/python2.6/site-packages/django/views/decorators/csrf.py", line 57, in wrapped_view
File: "/usr/lib/python2.6/site-packages/rest_framework/views.py", line 388, in dispatch
File: "/usr/lib/python2.6/site-packages/rest_framework/views.py", line 317, in initial
File: "/usr/lib/python2.6/site-packages/rest_framework/views.py", line 267, in perform_authentication
File: "/usr/lib/python2.6/site-packages/rest_framework/request.py", line 219, in user
File: "/usr/lib/python2.6/site-packages/rest_framework/request.py", line 385, in _authenticate
File: "/Locqus/API/PathServe/LocqusSimpleAuth.py", line 35, in authenticate
File: "/usr/lib/python2.6/site-packages/django/contrib/auth/__init__.py", line 49, in authenticate
File: "/usr/lib/python2.6/site-packages/django/contrib/auth/backends.py", line 16, in authenticate
File: "/usr/lib/python2.6/site-packages/django/contrib/auth/models.py", line 167, in get_by_natural_key
File: "/usr/lib/python2.6/site-packages/django/db/models/manager.py", line 151, in get
File: "/usr/lib/python2.6/site-packages/django/db/models/query.py", line 304, in get
File: "/usr/lib/python2.6/site-packages/django/db/models/query.py", line 77, in __len__
File: "/usr/lib/python2.6/site-packages/django/db/models/query.py", line 857, in _fetch_all
File: "/usr/lib/python2.6/site-packages/django/db/models/query.py", line 220, in iterator
File: "/usr/lib/python2.6/site-packages/django/db/models/sql/compiler.py", line 713, in results_iter
File: "/usr/lib/python2.6/site-packages/django/db/models/sql/compiler.py", line 785, in execute_sql
File: "/usr/lib/python2.6/site-packages/django/db/backends/__init__.py", line 162, in cursor
File: "/usr/lib/python2.6/site-packages/django/db/backends/__init__.py", line 132, in _cursor
File: "/usr/lib/python2.6/site-packages/django/db/backends/__init__.py", line 127, in ensure_connection
File: "/usr/lib/python2.6/site-packages/django/db/backends/__init__.py", line 115, in connect
File: "/usr/lib/python2.6/site-packages/django/db/backends/postgresql_psycopg2/base.py", line 115, in get_new_connect\
ion
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/hooks/database_dbapi2.py", line 91, in __call__
File: "/usr/local/lib64/python2.6/site-packages/psycopg2/__init__.py", line 164, in connect
conn = _connect(dsn, connection_factory=connection_factory, async=async)
其中一个留下了这个痕迹。
# ProcessId: 10353
# ThreadID: 140478641071872
File: "/usr/lib64/python2.6/threading.py", line 504, in __bootstrap
File: "/usr/lib64/python2.6/threading.py", line 532, in __bootstrap_inner
File: "/usr/lib64/python2.6/threading.py", line 484, in run
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/core/agent.py", line 543, in _harvest_loop
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/core/agent.py", line 584, in _run_harvest
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/core/application.py", line 1258, in harvest
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/core/internal_metrics.py", line 82, in __call__
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/core/data_collector.py", line 614, in send_metr\
ic_data
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/core/data_collector.py", line 340, in send_requ\
est
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/sessions.py", line 399, in po\
st
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/sessions.py", line 357, in re\
quest
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/sessions.py", line 460, in se\
nd
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/adapters.py", line 320, in se\
nd
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/packages/urllib3/connectionpo\
ol.py", line 544, in urlopen
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/packages/urllib3/connectionpo\
ol.py", line 369, in _make_request
File: "/usr/lib64/python2.6/httplib.py", line 920, in request
File: "/usr/lib64/python2.6/httplib.py", line 957, in _send_request
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/hooks/external_httplib.py", line 35, in httplib\
_endheaders_wrapper
File: "/usr/lib64/python2.6/httplib.py", line 914, in endheaders
File: "/usr/lib64/python2.6/httplib.py", line 786, in _send_output
File: "/usr/lib64/python2.6/httplib.py", line 745, in send
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/packages/urllib3/connectionpo\
ol.py", line 134, in connect
File: "/usr/lib64/python2.6/site-packages/newrelic-2.20.0.17/newrelic/packages/requests/packages/urllib3/util.py", li\
ne 626, in ssl_wrap_socket
File: "/usr/lib64/python2.6/ssl.py", line 338, in wrap_socket
File: "/usr/lib64/python2.6/ssl.py", line 118, in __init__
cert_reqs, ssl_version, ca_certs)
最后，我在 gdb 中附加了该进程并获取了线程信息
Id Target Id Frame
14 Thread 0x7fc3d695d700 (LWP 10355) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
13 Thread 0x7fc3d615c700 (LWP 10356) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
12 Thread 0x7fc3d595b700 (LWP 10357) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
11 Thread 0x7fc3d515a700 (LWP 10358) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
10 Thread 0x7fc3d4959700 (LWP 10359) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
9 Thread 0x7fc3cffff700 (LWP 10360) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
8 Thread 0x7fc3cf7fe700 (LWP 10361) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
7 Thread 0x7fc3ceffd700 (LWP 10362) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
6 Thread 0x7fc3ce7fc700 (LWP 10363) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
5 Thread 0x7fc3cdffb700 (LWP 10364) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
4 Thread 0x7fc3cd7fa700 (LWP 10365) "httpd.worker" 0x00007fc3e27e3705 in pthread_cond_wait@@GLIBC_2.3.2 ()
from /lib64/libpthread.so.0
3 Thread 0x7fc3bbfff700 (LWP 10372) "httpd.worker" 0x00007fc3e2300933 in select () from /lib64/libc.so.6
2 Thread 0x7fc3bb7fe700 (LWP 10373) "httpd.worker" 0x00007fc3e27e5f7d in __lll_lock_wait ()
from /lib64/libpthread.so.0
* 1 Thread 0x7fc3e3fa5840 (LWP 10353) "httpd.worker" 0x00007fc3e27e625d in read () from /lib64/libpthread.so.0
答案1
您所看到的看起来像是 Postgres 客户端库中的一个已知错误。
当存在并发线程并且您获得与 Web 服务等的 SSL 连接重叠的 SSL Postgres 连接时，Postgres 客户端库可能会导致死锁。
我不确定 Postgres 何时/是否会被修复，但最近有人针对 psycopg2 提出了一个解决方法。psycopg2 的问题是：