A
Aryeh Leib Taurog
I'm encountering the following error on my fastcgi web server and
would
greatly appreciate ANY pointers for debugging/fixing this problem.
*** glibc detected *** /usr/bin/python2.5: double free or corruption
(fasttop): 0x08b47d60 ***
If I don't set MALLOC_CHECK_ then the server just hangs and the above
message appears in the fastcgi error log. With MALLOC_CHECK_=0 I get
an
error message in the browser instead.
I'm using the following components (on a shared hosting account):
Debian lenny 64 bit
Python 2.5.2
lighttpd 1.4.19
Django 1.1.1
latest flup (today's hg tip, flup-server-ae5fe54fba18)
postgresql 8.3.9
psycopg2 2.0.13
What's interesting about the problem is it seems to happen only when
new
fcgi processes are spawned and only when a db query is made. That is,
the
frequency of the problem corresponds exactly with the maxrequests
setting
on my fcgi process. If I set maxrequests to 1 then I get this error
for
each request that hits the database. If I set it to 3 then I get the
error
with every third request, if that request hits the database.
If I put Django into DEBUG and set MALLOC_CHECK_=0 then I get the
following
error with traceback:
Exception Type: OperationalError
Exception Value: server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
Exception Location: /home/altaurog/django/db/backends/util.py in
execute, line 19
Regardless of which url/view I request, the exception always is raised
on a
sql execute statement.
would
greatly appreciate ANY pointers for debugging/fixing this problem.
*** glibc detected *** /usr/bin/python2.5: double free or corruption
(fasttop): 0x08b47d60 ***
If I don't set MALLOC_CHECK_ then the server just hangs and the above
message appears in the fastcgi error log. With MALLOC_CHECK_=0 I get
an
error message in the browser instead.
I'm using the following components (on a shared hosting account):
Debian lenny 64 bit
Python 2.5.2
lighttpd 1.4.19
Django 1.1.1
latest flup (today's hg tip, flup-server-ae5fe54fba18)
postgresql 8.3.9
psycopg2 2.0.13
What's interesting about the problem is it seems to happen only when
new
fcgi processes are spawned and only when a db query is made. That is,
the
frequency of the problem corresponds exactly with the maxrequests
setting
on my fcgi process. If I set maxrequests to 1 then I get this error
for
each request that hits the database. If I set it to 3 then I get the
error
with every third request, if that request hits the database.
If I put Django into DEBUG and set MALLOC_CHECK_=0 then I get the
following
error with traceback:
Exception Type: OperationalError
Exception Value: server closed the connection unexpectedly
This probably means the server terminated abnormally
before or while processing the request.
Exception Location: /home/altaurog/django/db/backends/util.py in
execute, line 19
Regardless of which url/view I request, the exception always is raised
on a
sql execute statement.