Morning,
I still haven't had any luck addressing this problem; I'm posting this
in the hope that someone has some sort of idea, otherwise I will have
to scrap mod_fcgid and revert to mod_php this evening, as this is
causing considerable problems for our users.
Thanks for any assistance,
Liam
I don't have any solutions for you, but I'd recommend trying to discover where
or why it's hanging. The tool 'strace' would be most useful in figuring this
out.
The ideas that come to mind off the top of my head are problems with file size,
the memory of the child processes, or something like
Hi Liam,
Try adding this to the VirtualHost configuration for the intended website:
IPCCommTimeout 300
In addition, I'd recommend starting with a bare-bones Apache (without any
additional 3rd party module, like Suhosin and others) and then activating
only FastCGI with PHP. Best is when you
Hi Josiah,
There was a strace snippet in the linked serverfault page, as follows
root@some-machine:~# strace -p 24837
Process 24837 attached - interrupt to quit
write(3,
\5|A\313%\35\337\376\275\237\230\266\242\371\37YjzD\322\215\357\336:M\362P\335\242\214\341...,
17432
Just hanging on that,
Liam,
Someone else commented that you should increase the value of your
IPCCommTimeout setting; I'd recommend this as well, though I have to ask, as
these uploads taking multiple minutes to complete or do they hang fairly
quickly?
Josiah Purtlebaugh
josiah.purtleba...@gmail.com
+1 253 797
What do you mean by RAM is almost full. Under normal usage the OS
will allocate most RAM not used by active processes to disk cache
leaving only a small amount free.
What is the load average and CPU idle percent when it gets bogged down.
Are your apps using a lot of disk resources?
What is the
Walter Heck walterheck at gmail.com writes:
Hi list,
I'm running:
[09-02-11 15:07:09] [root at server02.yomamma.com]
[~] # dpkg -l | egrep '{fcgi|php|apache|http}'
ii apache2 2.2.16-4~bpo50+1 Apache
HTTP Server metapackage
ii
Hi,
how can i get uploadprogress (http://pecl.php.net/package/uploadprogress)
working with mod_fcgid?
The progress bar gets no progress update.
Anyone the same problem?
Best regards
sunfire
--
uberSVN's rich system
Josiah Purtlebaugh josiah.purtlebaugh at gmail.com writes:
Hello,
From the page you linked:
It is only known to work on Apache with mod_php, other SAPI
implementations unfortunately still have issues.
At least PHP 5.2 is needed.
mod_php and mod_fcgid are fundamentally different; as
Exact Oliver,
I have the exact same issue, have you managed to solve the problem?
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed
Hello
I'm only getting started with FCGI to run Python web applications.
On a shared host with Apache + mod_fcgid + the Flup FastCGI server
that takes care of translating between FastCGI and WSGI, I noticed
that there's a several minute delay between when I edit a script and
when the change
Hello Gilles,
Are you familiar with how mod_fcgid executes scripts? Basically, the
executable is run and stays in memory so that it can serve any future
requests without forking additional processes per request. Once the
process reaches the timeout defined in the mod_fcgid configuration (or
the
On Fri, 28 Sep 2012 16:56:39 -0700, Josiah Purtlebaugh
josiah.purtleba...@gmail.com wrote:
Unfortunately, without access to the mod_fcgid configuration you are
at the mercy of the timeout or max requests variables defined by the
administrator. That is why it appears to take so long for your
Unfortunately, without access to the mod_fcgid configuration you are
at the mercy of the timeout or max requests variables defined by the
administrator. That is why it appears to take so long for your scripts
to update.
This is not exactly true. If you have an access to server's shell via ssh
On Sat, 29 Sep 2012 11:41:06 -0700, Josiah Purtlebaugh
josiah.purtleba...@gmail.com wrote:
Sure Anton, assuming that the mod_fcgid process is running as the same
user that he has shell access as.
ps aux | grep fcgi returns nothing. Can it run under a different
name?
ps aux | grep fcgi returns nothing. Can it run under a different
name?
You need to run ps axu | grep $(id -un) instead
--
Best regards
Anton Khalikov
--
Got visibility?
Most devs has no idea what their production app
On Mon, 1 Oct 2012 09:08:39 +0600, Anton Khalikov an...@khalikov.ru
wrote:
ps aux | grep fcgi returns nothing. Can it run under a different
name?
You need to run ps axu | grep $(id -un) instead
Thank you.
--
Everyone
Hello Group,
the problem using ITK+mod_fcgid is the fcgid socket. the socket is
owned by wwwuser (wwwrun / www-data) of apache.
ITK set vhost to user1, so fcgid says:
Permission denied: mod_fcgid: can't lock process table in pid
socket dir:
srwx-- 1 wwwrun wwwrun 0 21. Feb 14:11
No promises this will work as I am not a maintainer nor developer of
mod_fcgid, but I have prepared a patch for you that should achieve the same
effect. I'm not able to test compilation with mpm-itk (mainly because I'm
not sure which package you're using), but it compiles cleanly without
mpm-itk
Thanks, but it does not work:
srwx-- 1 wwwrun wwwrun 0 23. Feb 02:26 10055.6
error_log: (13)Permission denied: mod_fcgid: can't lock process table in pid
patch and compile without error
greets martin
Von: Josiah Purtlebaugh [mailto:josiah.purtleba...@gmail.com]
I'm sorry, I should have considered that an unprivileged user would be
unable to change ownership of an existing socket created with a different
uid. I'll have to consider further what additional changes might be
necessary.
Josiah
On Feb 22, 2013 5:34 PM, m...@unimx.de wrote:
Thanks, but it
We test:
/* Correct the file owner */
if (!geteuid()) {
//#if defined(MPM_ITK)
//if (chown(unix_addr.sun_path, procnode-uid, -1) 0) {
if (chown(unix_addr.sun_path, web589, -1) 0) {
//#else
//if (chown(unix_addr.sun_path, ap_unixd_config.user_id, -1) 0)
Right. The Unix socket is already owned by wwwrun and the chown is being
executed by a different uid, I think. We need to enforce not only the chown
but also the creation of the socket. It will be in another function.
Josiah
On Feb 22, 2013 5:48 PM, m...@unimx.de wrote:
We test:
** **
Yes the socket -i think- is created by apache with user root, only the child
get wwwrun and his child with mpm-itk get f.e. web589
So if we find the function create the socket, there it have to be created
with the user of mpm-itk.
May at this time the user is unknown.
May it is possible
101 - 124 of 124 matches
Mail list logo