apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
Hi, I use Pylons 1.0 Framework and mod_wsgi on production server. I can use my application but randomly 1 request on 10 I've error : Internal Server Error in browser and in error log file : [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/ I've read already many documentation page about this issue on mod_wsgi and Pylons website. Have you some idea about this error ? More information about my configuration : * I use apache2 from Debian package * I've compiled Python 2.6 from source code * I've compiled mod_wsgi-3.3 from source code (I've also tested with mod_wsgi-2.8, same error) VirtualHost *:80 ServerName example.com IfModule mod_wsgi.c WSGIScriptAlias / /home/examplecom/demo/demo.wsgi WSGIDaemonProcess mydemo threads=1 maximum-requests=500 inactivity-timeout=500 processes=1 display-name=%{GROUP} WSGIProcessGroup mydemo WSGIApplicationGroup %{GLOBAL} /IfModule ErrorLog /home/examplecom/demo/error.log LogLevel info /VirtualHost Thanks for your help. Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: [modwsgi] apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
On 8 December 2010 20:37, Stéphane Klein steph...@harobed.org wrote: Hi, I use Pylons 1.0 Framework and mod_wsgi on production server. I can use my application but randomly 1 request on 10 I've error : Internal Server Error in browser and in error log file : [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/ Need to see the error messages before and after this and for matching time in main Apache error log if you are using a virtual host specific error log. This message will arise in a few circumstances. The first is that WSGI script file is being executed as a CGI script for some reason. The second is that the mod_wsgi daemon process is crashing. The third is that daemon process is being forcibly restarted while there are active requests which haven't responded in time. I've read already many documentation page about this issue on mod_wsgi and Pylons website. Have you some idea about this error ? More information about my configuration : * I use apache2 from Debian package * I've compiled Python 2.6 from source code * I've compiled mod_wsgi-3.3 from source code (I've also tested with mod_wsgi-2.8, same error) VirtualHost *:80 ServerName example.com IfModule mod_wsgi.c WSGIScriptAlias / /home/examplecom/demo/demo.wsgi WSGIDaemonProcess mydemo threads=1 maximum-requests=500 inactivity-timeout=500 processes=1 display-name=%{GROUP} A very odd configuration. You are running a single process with a single thread in it. That it is a single thread only though, means that you can't be incurring third possibility above though as a forced restart due to maximum requests being reached as there can't be a concurrent request that is being interrupted. You can get a forced shutdown when request is still current due to inactivity timeout if the request hangs for 500 seconds for your configuration, but in that case I would say you would notice the hang given that that is almost 10 minutes before that could kick in. I would suggest so as to get more information if it is something related to maximum requests or inactivity timeout, that you ensure using: LogLevel info in Apache configuration and not the default: LogLevel warn Have a good look for segmentation fault messages in main Apache error log as well. Graham WSGIProcessGroup mydemo WSGIApplicationGroup %{GLOBAL} /IfModule ErrorLog /home/examplecom/demo/error.log LogLevel info /VirtualHost Thanks for your help. Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups modwsgi group. To post to this group, send email to modw...@googlegroups.com. To unsubscribe from this group, send email to modwsgi+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/modwsgi?hl=en. -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
Le 08/12/2010 10:54, Graham Dumpleton a écrit : On 8 December 2010 20:37, Stéphane Kleinsteph...@harobed.org wrote: Hi, I use Pylons 1.0 Framework and mod_wsgi on production server. I can use my application but randomly 1 request on 10 I've error : Internal Server Error in browser and in error log file : [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/ Need to see the error messages before and after this and for matching time in main Apache error log if you are using a virtual host specific error log. This message will arise in a few circumstances. This is more information messages : http://pastebin.com/myi5D5Xn Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
Le 08/12/2010 11:21, Stéphane Klein a écrit : Le 08/12/2010 10:54, Graham Dumpleton a écrit : On 8 December 2010 20:37, Stéphane Kleinsteph...@harobed.org wrote: Hi, I use Pylons 1.0 Framework and mod_wsgi on production server. I can use my application but randomly 1 request on 10 I've error : Internal Server Error in browser and in error log file : [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/ Need to see the error messages before and after this and for matching time in main Apache error log if you are using a virtual host specific error log. This message will arise in a few circumstances. This is more information messages : http://pastebin.com/myi5D5Xn Well, I see the last message of this thread : http://groups.google.com/group/modwsgi/browse_thread/thread/ba82b2643564d2dd He say : Last night we installed mod_wsgi 3.3, and the KeyErrors have totally disappeared. I highly recommend the 3.3 upgrade if you're having this problem. Thanks, Graham. In my test, I've tested mod_wsgi 3.3 and 2.8 In last log message, it's 2.8 version. Now I've reinstalled 3.3 version, this is the new log messages : http://pastebin.com/kKztWiCY Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
Logs show the segmentation faults I speak of: [Wed Dec 08 11:32:47 2010] [error] [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/employees/ [Wed Dec 08 11:32:47 2010] [notice] child pid 22620 exit signal Segmentation fault (11) Attach gdb to daemon process as described in: http://code.google.com/p/modwsgi/wiki/DebuggingTechniques#Debugging_Crashes_With_GDB enter 'cont' command so runs again and then keep firing off requests until it crashes. Suggest removing maximum-requests during that time so that process isn't restarted because of that number being exceeded. In other words, debug why process is crashing. For ideas of why, see: http://code.google.com/p/modwsgi/wiki/FrequentlyAskedQuestions#Apache_Process_Crashes http://code.google.com/p/modwsgi/wiki/ApplicationIssues Graham On 8 December 2010 21:37, Stéphane Klein steph...@harobed.org wrote: Le 08/12/2010 11:21, Stéphane Klein a écrit : Le 08/12/2010 10:54, Graham Dumpleton a écrit : On 8 December 2010 20:37, Stéphane Kleinsteph...@harobed.org wrote: Hi, I use Pylons 1.0 Framework and mod_wsgi on production server. I can use my application but randomly 1 request on 10 I've error : Internal Server Error in browser and in error log file : [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/ Need to see the error messages before and after this and for matching time in main Apache error log if you are using a virtual host specific error log. This message will arise in a few circumstances. This is more information messages : http://pastebin.com/myi5D5Xn Well, I see the last message of this thread : http://groups.google.com/group/modwsgi/browse_thread/thread/ba82b2643564d2dd He say : Last night we installed mod_wsgi 3.3, and the KeyErrors have totally disappeared. I highly recommend the 3.3 upgrade if you're having this problem. Thanks, Graham. In my test, I've tested mod_wsgi 3.3 and 2.8 In last log message, it's 2.8 version. Now I've reinstalled 3.3 version, this is the new log messages : http://pastebin.com/kKztWiCY Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
Le 08/12/2010 11:52, Graham Dumpleton a écrit : Logs show the segmentation faults I speak of: [Wed Dec 08 11:32:47 2010] [error] [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/employees/ [Wed Dec 08 11:32:47 2010] [notice] child pid 22620 exit signal Segmentation fault (11) Attach gdb to daemon process as described in: http://code.google.com/p/modwsgi/wiki/DebuggingTechniques#Debugging_Crashes_With_GDB enter 'cont' command so runs again and then keep firing off requests until it crashes. This is the debug message : http://pastebin.com/q9Lb72vQ Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
No clues unfortunately, just looks like arbitrary memory corruption. Do make sure you aren't loading mod_python into same Apache. Other than that, looks like a C extension module may be corrupting memory. What C extension modules for Python are you using? Graham On 8 December 2010 22:14, Stéphane Klein steph...@harobed.org wrote: Le 08/12/2010 11:52, Graham Dumpleton a écrit : Logs show the segmentation faults I speak of: [Wed Dec 08 11:32:47 2010] [error] [client 82.231.115.1] Premature end of script headers: demo.wsgi, referer: http://example.com/employees/ [Wed Dec 08 11:32:47 2010] [notice] child pid 22620 exit signal Segmentation fault (11) Attach gdb to daemon process as described in: http://code.google.com/p/modwsgi/wiki/DebuggingTechniques#Debugging_Crashes_With_GDB enter 'cont' command so runs again and then keep firing off requests until it crashes. This is the debug message : http://pastebin.com/q9Lb72vQ Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
Le 08/12/2010 12:20, Graham Dumpleton a écrit : No clues unfortunately, just looks like arbitrary memory corruption. Do make sure you aren't loading mod_python into same Apache. No, mod_python isn't installed. Other than that, looks like a C extension module may be corrupting memory. What C extension modules for Python are you using? this is more information about library loading found with gdb : http://pastebin.com/VFPe5BcK There are some clues in this log ? Regards, Stephane -- Stéphane Klein steph...@harobed.org - French blog: http://stephane-klein.info twitter: http://twitter.com/klein_stephane pro: http://www.is-webdesign.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: webhelpers.paginate GenerationError with url_for() in routes
On Tue, Dec 7, 2010 at 11:08 PM, Joel jrwdun...@gmail.com wrote: Ok, so I used grep to search for WebHelpers[=] as Mike suggested and found: /path/to/virtualenv/lib/python2.5/site-packages/FormBuild-2.2.0- py2.5.egg/EGG-INFO/requires.txt:WebHelpers=0.6.1,0.6.99 Argh, that could trap other people too. Thanks for letting us know. I still don't understand why url_for() didn't work, but I'm glad your application is running again. -- Mike Orr sluggos...@gmail.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Weird key problems with beaker_cache (cache depends on domain)
Hi, I'm using beaker_cache to cache results of some actions. The problem is, their result (the template used) depends sometimes on the domain which is used in the request (the same application is available under different domains). Right now I have something like this: @beaker_cache(expire=600, query_args=True) def group(self, id, portal=request.environ.get('HTTP_HOST', '')): The portal argument is there just to force beaker_cache to include the domain in key generation (I didn't find another way). This however, does not work as expected. I try to access the page first using domain1, then domain2. If I do raise Exception at the beginning of the action I can see that for domain1 key_dict = {'id': u'1', 'portal': 'domain1'}, but for domain2 key_dict is also {'id': u'1', 'portal': 'domain1'} ! Why is it like that? Is the request object somehow cached too? Why? How can I disable it? Regards, -- Juliusz Gonera http://juliuszgonera.com/ -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: Weird key problems with beaker_cache (cache depends on domain)
On Wed, Dec 08, 2010 at 08:29:36PM +0100, Juliusz Gonera wrote: I'm using beaker_cache to cache results of some actions. The problem is, their result (the template used) depends sometimes on the domain which is used in the request (the same application is available under different domains). Right now I have something like this: @beaker_cache(expire=600, query_args=True) def group(self, id, portal=request.environ.get('HTTP_HOST', '')): That's not going to work -- as you've discovered. The portal argument is there just to force beaker_cache to include the domain in key generation (I didn't find another way). This however, does not work as expected. I try to access the page first using domain1, then domain2. If I do raise Exception at the beginning of the action I can see that for domain1 key_dict = {'id': u'1', 'portal': 'domain1'}, but for domain2 key_dict is also {'id': u'1', 'portal': 'domain1'} ! Why is it like that? The default value for function arguments is computed during import time. Pylons imports the controllers dynamically during request processing, and that's the only reason it even works -- but it uses the HTTP_HOST value of the first request that triggered the import. I don't know the correct solution; you might need to write your own cache decorator. http://pypi.python.org/pypi/decorator is useful for those. Marius Gedminas -- We don't really understand it, so we'll give it to the programmers. signature.asc Description: Digital signature
Pyramid Pattern Request: Schema Migrations
Inspired by http://blog.cherouvim.com/a-table-that-should-exist-in-all-projects-with-a-database/ What's a good way to manage schema migrations in a Pyramid project? SQLAlchemy Migrate is a promising tool for RDBMSes. What would be a nice way to integrate it? Nice, in my book, means things like having paster serve myapp.ini check on startup if the database schema is recent enough and either perform the necessary migrations automatically -- if the .ini file allows this -- or abort and tell me what commands to run to perform the migrations manually. Last time I tried SQLALchemy Migrate on a Pylons project, I hit a fog of confusion (what migrate commands do I need to run when, exactly?), and then, eventually, a brick wall (in the sense of various changes being unsupported with sqlite, which I wanted to use during development, for simplicity). I ended up filing a bunch of issues and reinventing my own bicycle. I see that sqlite support is much improved nowadays. Similarly, when you're using ZODB with Pyramid, is there something like zope.generations for it? Marius Gedminas -- It's my understanding that although in principle TCP can handle huge throughputs in practice many stacks haven't been optimized for that case, so you have to either use a utility which opens multiple TCP sessions in parallel or do something really radical like upgrade to the latest version of the linux kernel. -- Bram Cohen signature.asc Description: Digital signature
Re: apache2 lenny pylons mod_wsgi python 2.6 = error : Premature end of script headers: demo.wsgi, referer: http://example.com/
On 9 December 2010 00:34, Stéphane Klein steph...@harobed.org wrote: Le 08/12/2010 12:20, Graham Dumpleton a écrit : No clues unfortunately, just looks like arbitrary memory corruption. Do make sure you aren't loading mod_python into same Apache. No, mod_python isn't installed. Other than that, looks like a C extension module may be corrupting memory. What C extension modules for Python are you using? this is more information about library loading found with gdb : http://pastebin.com/VFPe5BcK There are some clues in this log ? Only notable things are: Reading symbols from /usr/lib/libpq.so.5... (no debugging symbols found)...done. Loaded symbols for /usr/lib/libpq.so.5 Reading symbols from /usr/lib/libmysqlclient_r.so.15...(no debugging symbols found)...done. Loaded symbols for /usr/lib/libmysqlclient_r.so.15 Reading symbols from /usr/lib/libsqlite3.so.0...(no debugging symbols found)...done. Loaded symbols for /usr/lib/libsqlite3.so.0 Reading symbols from /usr/lib/libexpat.so.1...(no debugging symbols found)...done. Loaded symbols for /usr/lib/libexpat.so.1 Reading symbols from /usr/lib/libxml2.so.2...(no debugging symbols found)...done. Loaded symbols for /usr/lib/libxml2.so.2 These are preloaded by Apache or PHP extensions. If you are using any C extension module for Python which is expecting different versions of these libraries, you can end up with problems. If you can, don't load PHP into Apache and see if your problem goes away. Recommended that you don't use mod_php and mod_wsgi in same Apache if you can help it. Graham -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: Pyramid Pattern Request: Schema Migrations
On Thu, 2010-12-09 at 03:12 +0200, Marius Gedminas wrote: Inspired by http://blog.cherouvim.com/a-table-that-should-exist-in-all-projects-with-a-database/ What's a good way to manage schema migrations in a Pyramid project? Daniel may murder me for mentioning it (I don't know if he intended it for public consumption), but he wrote this: http://bitbucket.org/dholth/stucco_evolution/src It is a plugin to http://docs.repoze.org/evolution/ for SQLAlchemy. SQLAlchemy Migrate is a promising tool for RDBMSes. What would be a nice way to integrate it? Nice, in my book, means things like having paster serve myapp.ini check on startup if the database schema is recent enough and either perform the necessary migrations automatically -- if the .ini file allows this -- or abort and tell me what commands to run to perform the migrations manually. Last time I tried SQLALchemy Migrate on a Pylons project, I hit a fog of confusion (what migrate commands do I need to run when, exactly?), and then, eventually, a brick wall (in the sense of various changes being unsupported with sqlite, which I wanted to use during development, for simplicity). I ended up filing a bunch of issues and reinventing my own bicycle. I see that sqlite support is much improved nowadays. Similarly, when you're using ZODB with Pyramid, is there something like zope.generations for it? repoze.evolution is basically the guts of zope.generation. - C -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Routes ur() issue with URL
I sometimes use Route's url() to do the query args for me. For instance, url('http://google.com', q='aasdf') will give me: http://google.com?q=aasdf which is a very useful shortcut. However, I've found that this doesn't work: url('http://google.com?q=aasdf', foo='bar') which gives me: http://google.com?q=aasdf?foo=bar Does it make sense to modify Routes to parse the query before it applies the args in this special case? I can submit a patch. -- Jonathan Gardner jgard...@jonathangardner.net -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.
Re: Routes ur() issue with URL
On Wed, Dec 8, 2010 at 10:23 PM, Jonathan Gardner jgard...@jonathangardner.net wrote: I sometimes use Route's url() to do the query args for me. For instance, url('http://google.com', q='aasdf') will give me: http://google.com?q=aasdf which is a very useful shortcut. However, I've found that this doesn't work: url('http://google.com?q=aasdf', foo='bar') which gives me: http://google.com?q=aasdf?foo=bar Does it make sense to modify Routes to parse the query before it applies the args in this special case? I can submit a patch. Use ``webhelpers.util.update_params()`` to add/modify query parameters on a URL: update_params(http://google.com/?q=aasdf;, foo=bar) update_params(http://google.com/;, q=aasdf, foo=bar) BTW, a HTTP URL without a slash after the domain is technically invalid. The browser or library may add it for you, but it's better to publish a complete URL. -- Mike Orr sluggos...@gmail.com -- You received this message because you are subscribed to the Google Groups pylons-discuss group. To post to this group, send email to pylons-disc...@googlegroups.com. To unsubscribe from this group, send email to pylons-discuss+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/pylons-discuss?hl=en.