ok LightDot,
just to confirm the tweak i should do for the sake of stability. you are
saying that under each VirtualHost i should have a separate
WSGIDaemonProcess under it. or, i should remove the WSGIDaemonProcess from
the wsgi.conf file and place it under each domains VirtualHost section.
Ok, so you're using one mod_wsgi / web2py instance and map virtualhosts
into specific apps within. Looks like a neat solution, but it seems to be
problematic in the same way Leonel is describing for the Windows
environment. Basically, different virtualhosts under the same
WSGIDaemonProcess have
ok, so centos 6x is the way to go for the most stable, reliable, etc., in
excellence of computing? that is what you are saying, because i really
love centos and have been using it for years. but i have been doubting my
httpd.conf setups for a while because all of those settings and then add in
Apache and mod_wsgi under CentOS 6.x (or any other RHEL derivative, like
Scientific Linux) tick like a clock. We us this combo for web2py almost
everywhere.
Are you using mod_wsgi as a deamon? Have you set up different process
groups, users etc. for different virtual hosts? You could post the r
oh yeah, i forgot to mention. i am running multiple domain names and i am
using VirtualHost under apache to redirect to the proper web2py
application. could that be the problem? i am using the session as files,
i like that better then the DB option because, well, i just like it because
it is
No it's not wrong to use apache and wsgi, although there are better options
if you use Linux, in my experience there's only a problem if you're running
it on ms windows and you have more than one web2py instance as they share
the same embedded mod_wsgi.
--
Resources:
- http://web2py.com
- http
sorry i dropped off the map for a while. my server had a hard hardware
crash and i have been managing its replacement. i think the processor or
motherboard went bad because the memory and HD are testing out ok. i
actually decided to goto amazon's EC2 and i am using centos 6,4 under that.
it
My suspicion right now is that this bug is indeed caused by mod_wsgi, and
it's not even solvable in windows where you don't have WSGIDaemonProcess,
so that's why restarting apache temporarily solves the problem, what I'm
going to do is migrate the server to Linux (including migrating the only
p
On my tests, this is happening after I go to the admin for other
applications on the same server, so it's possible that admin is the one
causing the problem.
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code
I don't think I can send you the file as this crash happens precisely when
it's trying to store the cookie and response.session_id is None.
I have no idea how thread.lock objects are getting into the session.
I have tried to solve this myself, but the code for Session is not exactly
easy to
t would help a lot to know what the session object looks like so that we
can try reproduce the problem. I assume this is for a particular session
object. Can you retrieve the session_id from the cookie and send me the
corresponding file? Do you know how the lock object gets into the session?
I
I do not fully understand the details but there is a note in the pyhton
docs that says that __getstate__ can return False. If it does __setstate__
is not called.
If I change __getstate__ to return dict(self) or if I remove it, sessions
break.
On Tuesday, 3 December 2013 13:50:33 UTC-6, Jonathan
I'm still getting another problem if I store cookies in files as usual:
Traceback (most recent call last):
File "C:\blueweb2py\gluon\main.py", line 479, in wsgibase
session._try_store_in_cookie_or_file(request, response)
File "C:\blueweb2py\gluon\globals.py", line 1089, in
_try_store_in_c
Is this the problem for everybody else in the thread?
On Monday, 16 December 2013 12:44:54 UTC-6, Leonel Câmara wrote:
>
> After further investigation I've found what triggers this problem for me.
>
> If you have several virtualhosts in apache, on the same IP:PORT, each
> running from their own w
After further investigation I've found what triggers this problem for me.
If you have several virtualhosts in apache, on the same IP:PORT, each
running from their own web2py folder, you can get this problem.
For instance, you only have 2 subdomains
foo.example.com
bar.example.com
for foo
I just found this:
https://groups.google.com/forum/#!searchin/web2py/weird$20auth$20error/web2py/jkqyPM5_zaE/Ve5gy7greuoJ
This is version 2.4.5 so it is NOT something introduced in recent version.
I am also privileged to feel the pain that this issue causes in production.
Marin
--
Resources:
-
Lucas, did the workaround of using the db to store the session work for you?
Terça-feira, 3 de Dezembro de 2013 11:59:11 UTC, lucas escreveu:
>
> hey everyone,
>
> thanx for the help on my prior posts, things are moving along. i now get
> a crash when i try to login under the application, the t
hey leonel,
i am still here. watching and waiting. no worries. i did try the
sessions delete and the pyc delete under gluon but the exception does
return rather immediately. i am not as good as tracing through code as
you, but it does seem like a hard one. sorry for being the messenger.
l
Ignore, my patch still sees other bugs, such as trying to pickle lock
objects, I have to examine this more carefully. This is a particularly
interesting bug.
In the mean time, as a workaround, I haven't been able (yet) to see the
issue using the db to store the cookies session.connect(request,
Of course we can always just remove __getstate__ = lambda self: None from
Storage
However I think that's there to force the use of save_storage which does
the proper file locking. I can understand Session not having the dict
conversion since it isn't needed by FastStorage but in the mean time w
I've found the problem, Session in globals.py constantly tries to pickle
itself forgetting to convert to a dict first. As it should if you look at
storage.py save_storage.
I guess I can make a patch for this.
Terça-feira, 3 de Dezembro de 2013 11:59:11 UTC, lucas escreveu:
>
> hey everyone,
>
Never mind my previous response, I'm getting the exact same error on apache
using mod_wsgi on microsoft windows 2012. I guess I will try to find out
what's wrong because this is really annoying. The error seems to go away if
you delete sessions, but it comes back seemingly randomly.
Terça-feira
Lucas I kept getting this problem too, I found that my problem was old .pyc
files in my web2py/gluon directory. If you have the same problem delete
them all and restart apache.
Terça-feira, 3 de Dezembro de 2013 11:59:11 UTC, lucas escreveu:
>
> hey everyone,
>
> thanx for the help on my prior p
does anyone else have this problem, or is it only me? is there at least
acknowledgement that is is a problem, or is my sanity not in check? should
i change my programming approach in some way or am i missing a handler or
loading of a handler? just waiting on some expert advice.
--
Resources
any ideas on this exception and trace? a little anxious about it because
my site is essentially offline. lucas
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list (Report Issue
On 3 Dec 2013, at 9:59 AM, Massimo Di Pierro wrote:
> Is this with 2.8.2?
This confuses me. In Storage, we say:
__getstate__ = lambda self: None
...but the pickling logic ends up calling what __getstate__ returns (without
checking for None), which would explain the crash below. The confus
yes it was/is with 2.8.2, but running only under apache, it is fine with
"python web2py" AND i did erase all session files, as in "rm -rf
sessions/*" which is a hard remove in linux. lucas
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/w
I had the same problem, I solved it by manually deleting all sessions, as
in deleting everything in the "sessions" folder of the application.
Terça-feira, 3 de Dezembro de 2013 11:59:11 UTC, lucas escreveu:
>
> hey everyone,
>
> thanx for the help on my prior posts, things are moving along. i no
Is this with 2.8.2?
On Tuesday, 3 December 2013 05:59:11 UTC-6, lucas wrote:
>
> hey everyone,
>
> thanx for the help on my prior posts, things are moving along. i now get
> a crash when i try to login under the application, the traceback is:
>
> Traceback (most recent call last):
> File "/op
oh, btw, the above does not happen under "python web2py.py" but only under
the apache umbrella. lucas
--
Resources:
- http://web2py.com
- http://web2py.com/book (Documentation)
- http://github.com/web2py/web2py (Source code)
- https://code.google.com/p/web2py/issues/list (Report Issues)
---
Yo
30 matches
Mail list logo
