I need to reload my php.ini and there’s nothing in the help dialog about restarting it.

Assuming you have the appropriate RC scripts in place,

/etc/init.d/php-fpm restart

or

restart php-fpm

or

service php-fpm restart

or

service php5-fpm restart

Or whatever the equivalent is on your system.

Edited to include suggestion from Matt Winckler.

Check more discussion of this question.

I have these rewrite rules (I tried both with no avail):

location ~* "^/([a-z0-9]{32})\.png$" {
  rewrite ^ /index.php?page=log&id=$1 last;
} 

and

location ~* "/(?<hash>[a-z0-9]{32})\.png" {
  rewrite ^ /index.php?page=log&id=$hash;
}

and

location / {
  try_files $uri $uri/ @rewrites;
}location @rewrites {
  rewrite "^/([a-zA-Z0-9]{32})\.png$" /index.php?page=log&id=$1 last;
  #...
}

Basically, I want the URL http://example.com/aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa.png to actually pass the parameters to my index.php script, and in the script, I got:

$db->save_hash($_GET['id']);
header('Content-type: image/png');
readfile('images/beacon.png');
break;

But nginx is giving me a “not found”, but other rewrites work fine. What gives?

As per nginx documentation of how the location directive works:

A location can either be defined by a prefix string, or by a regular expression.
…
A search of regular expressions terminates on the first match, and the corresponding configuration is used.

Based on addition information from your own answer to this very question, it would seem like you intend on having several location directives to apply simultaneously right at the very same time, which is explicitly not allowed as per the unambiguous documentation.

Remember, nginx is faster and cleaner by design. So, don’t expect the quirks!

Check more discussion of this question.

I have a html presentation which I would like to serve on a subdomain on my server. It just consists of some static html files and some assets (the Github is here).

I’m trying to write an nginx config file which will serve the presentation. What I have so far is the following:

server {
  server_name make-js-like-ruby.grinnick.com;  root /home/dtuite/stuff/presentations/make_js_like_ruby/introduction;  access_log off;
  error_log off;
}

The actual root file is in /introduction/index.html. Nginx seems to serve that file by default which is fine. The problem is that none of the assets seems to be available on the page. You can see the problem live here..

How do I make the assets are available but also make sure that the subdomain roots to the index file correctly?

Move Root Path

Your current directory structure (on your hard drive) is like following

/home/dtuite/stuff/presentations/make_js_like_ruby/
    code/
    core/
    extensions/
    images/
    introduction/
    themes/

However, your root path is set to

/home/dtuite/stuff/presentations/make_js_like_ruby/introduction

The web server will not access any file outside that path. Since themes, images, etc., are all outside the above path, their content will not be served.

Fix

I suggest moving the root path one level up

root /home/dtuite/stuff/presentations/make_js_like_ruby;

Then move the index.html into /home/dtuite/stuff/presentations/make_js_like_ruby. Of course links within index.html will have to be modified accordingly.

Check more discussion of this question.

I’ve made some changes in the .ini files and wanted to restart php5-fpm.

/etc/init.d/php5-fpm restart

Jan 20 14:25:48.171577 [ERROR] bind() for address
‘/var/lib/php5-fpm/apps.sock’ failed: Address already in use (98)

How can I find out what is using this address? I’ve already stopped nginx but that didn’t change anything.

You can use lsof

lsof /var/lib/php5-fpm/apps.sock

The output is lines containing the information

COMMAND, PID, USER, FD, TYPE, DEVICE, SIZE/OFF, NODE, NAME e.g.

lsof /lib64/libnss_files-2.12.so
COMMAND  PID USER  FD   TYPE DEVICE SIZE/OFF   NODE NAME
bash    2240 iain mem    REG  253,0    65928 142129 /lib64/libnss_files-2.12.so
bash    4700 iain mem    REG  253,0    65928 142129 /lib64/libnss_files-2.12.so

Check more discussion of this question.

I’m using http public cache all over an images website to allow clients and anyone between the clients and my server (eg: cloudflare, isp, etc) to cache HTML, assets (js/css/icons), and images (the main content of the website). But I don’t want nginx to cache these images, because it will take too much space. I’m not sure whether it’s happening, I don’t know how to check.

So how do I check, and if it’s happening, how do I stop it, without also stopping client and proxies cache?

Note: I’m using nginx as web server with fastcgi, and not as a proxy

If you haven’t explicitly set up nginx as a caching proxy, then it will not cache your images or other static content, but merely serve it from the filesystem.

Check more discussion of this question.

In nginx.conf:

http { 
    geoip_country /etc/nginx/GeoIP.dat;
    ...
}

If I do:

server{
    ...
    location / {
        add_header X-Geo $geoip_country_code;                   
        add_header X-Geo3 $geoip_country_code3;
        add_header X-IP $remote_addr;
        ...
    }
}

Only X-IP show up in my headers.

$ curl -I www.example.org
HTTP/1.1 302 FOUND
Content-Type: text/html; charset=utf-8
Date: Thu, 17 Jan 2013 19:29:23 GMT
Location: http://www.example.org/login/?next=/
Server: nginx/1.2.2
Vary: Cookie
X-IP: 10.139.34.12
Connection: keep-alive

If I change the location block to:

location / {
    add_header X-Geo "foo";                   
    add_header X-Geo3 "bar";
    add_header X-IP $remote_addr;
    ...
}

The headers show up, how can I get $geoip_country_code?

Ah I just figure out what was happening. It was trying to resolve the GeoIP for the LoadBalancers IP rather than the client which is why it was blank. Once I curled the host directly, it returned the correct result. I had to use the HttpRealipModule to give the HttpGeoipModule the clients IP. In my conf:

real_ip_header X-Forwarded-For;
set_real_ip_from 0.0.0.0/0;

Check more discussion of this question.

Server run on Amazon Linux AMI** (similary to redhat)

I follow the installation from Install PHP-FPM, Nginx, & MySQL on EC2 with Amazon Linux AMI

It’s work properly if I call directly from server IP address like 174.129.213.152/index.php
and I can access other file from my domain correctly like http://gamebiz.asia/robots.txt

The problem is when I call php file http://gamebiz.asia or http://gamebiz.asia/index.php their result in blank page.

I try change configuration from

fastcgi_pass unix:/var/run/php-fpm/php-fpm.sock; to fastcgi_pass 127.0.0.1:9000; and fastcgi_pass localhost:9000;

the result when I call php file is “502 Bad Gateway” and If I remove fastcgi_pass I got “download” file instead of execute php file

Configuration

conf.d/default.conf

server{
    location ~* \.(jpg|jpeg|gif|png|css|js|ico|xml|mp3|ogg|wav|aac)$ {
        access_log        off;
        log_not_found     off;
        expires           max;
    }
    location ~* \.php$ {
        fastcgi_index   index.php;
        fastcgi_pass    unix:/var/run/php-fpm/php-fpm.sock;
        include /etc/nginx/fastcgi_params;
        fastcgi_param   SCRIPT_FILENAME    $document_root$fastcgi_script_name;
        fastcgi_param   SCRIPT_NAME        $fastcgi_script_name;
    }
    server_tokens off;
}

sites-available/gamebiz.asia.conf

server {
    server_name  www.gamebiz.asia;
    rewrite ^(.*) http://gamebiz.asia$1 permanent;
}server {
    listen 80;
    server_name gamebiz.asia;
    root   /var/www/html/SpotDiff/Release;
    index index.php index.html;    location ~ \.php$ {
        fastcgi_pass    unix:/var/run/php-fpm/php-fpm.sock;
        fastcgi_index index.php;
        include /etc/nginx/fastcgi_params;
    }
}

I try config virtual host and sub-domain follow Installing NGINX and PHP-FPM – Setup for NGINX

they said I need to setup the different UNIX Sockets for PHP-FPM in /etc/php5/fpm/pool.d/

Unfortunately Amazon Linux AMI don’t have apt-get and I don’t have directory /etc/php5/fpm/pool.d/ and I search for pool.d but don’t know how to get this file and directory.

I end up on change fastcgi_pass from unix:/var/run/php5-fpm/DOMAINNAME.socket; to my default unix:/var/run/php-fpm/php-fpm.sock; but still got blank page

Please help me I struggle with this for 3 whole days. I will appreciate any kind of help. Thank you.

Seems you are missing

 fastcgi_param   SCRIPT_FILENAME    $document_root$fastcgi_script_name;
 fastcgi_param   SCRIPT_NAME        $fastcgi_script_name;

from the config of server_name gamebiz.asia; config block.

Check more discussion of this question.

I am running an Ubuntu Server that is starting to have problems. It’s very lightly used but am confused to why it is now having memory issues. I’ve included some of the output from syslog and free -m. What is this telling me exactly and what should I do next? Googling has provided some ambigous answers. It is only a 512MB instance and is set with 512MB of swap with a memory limit of 512 MB.

thx in advance for any help

Jan 14 06:25:15 curren rsyslogd: [origin software="rsyslogd" swVersion="5.8.1" x-pid="2311" x-info="http://www.rsyslog.com"] rsyslogd was HUPed
Jan 14 06:25:16 curren CRON[3242]: (CRON) info (No MTA installed, discarding output)
Jan 14 06:40:59 curren kernel: mysqld invoked oom-killer: gfp_mask=0x201da, order=0, oom_adj=0, oom_score_adj=0
Jan 14 06:40:59 curren kernel: mysqld cpuset=/ mems_allowed=0
Jan 14 06:40:59 curren kernel: Pid: 2414, comm: mysqld Not tainted 3.6.5-x86_64-linode28 #1
Jan 14 06:40:59 curren kernel: Call Trace:
Jan 14 06:40:59 curren kernel: [<ffffffff810c5fd6>] dump_header+0x7d/0x1af
Jan 14 06:40:59 curren kernel: [<ffffffff810a2d96>] ? delayacct_end+0x74/0x7f
Jan 14 06:40:59 curren kernel: [<ffffffff816ef5a2>] ? _raw_spin_unlock_irqrestore+0x19/0x1c
Jan 14 06:40:59 curren kernel: [<ffffffff813ad18e>] ? ___ratelimit+0xe6/0x104
Jan 14 06:40:59 curren kernel: [<ffffffff810c6372>] oom_kill_process+0x66/0x30c
Jan 14 06:40:59 curren kernel: [<ffffffff810c6822>] out_of_memory+0x20a/0x263
Jan 14 06:40:59 curren kernel: [<ffffffff810ca096>] __alloc_pages_slowpath+0x403/0x4bc
Jan 14 06:40:59 curren kernel: [<ffffffff810ca299>] __alloc_pages_nodemask+0x14a/0x1ae
Jan 14 06:40:59 curren kernel: [<ffffffff813ac0b0>] ? radix_tree_lookup+0x1e/0x54
Jan 14 06:40:59 curren kernel: [<ffffffff810f6391>] alloc_pages_current+0xc5/0xe6
Jan 14 06:40:59 curren kernel: [<ffffffff810c3afb>] __page_cache_alloc+0x87/0x94
Jan 14 06:40:59 curren kernel: [<ffffffff810c51d4>] filemap_fault+0x25a/0x341
Jan 14 06:40:59 curren kernel: [<ffffffff810e1b69>] __do_fault+0xa5/0x3a0
Jan 14 06:40:59 curren kernel: [<ffffffff810e2d03>] handle_pte_fault+0xdd/0x1b5
Jan 14 06:40:59 curren kernel: [<ffffffff81004a4c>] ? xen_pmd_val+0x9/0xb
Jan 14 06:40:59 curren kernel: [<ffffffff81004659>] ? __raw_callee_save_xen_pmd_val+0x11/0x1e
Jan 14 06:40:59 curren kernel: [<ffffffff810e2f32>] handle_mm_fault+0x157/0x169
Jan 14 06:40:59 curren kernel: [<ffffffff816f2947>] do_page_fault+0x364/0x389
Jan 14 06:40:59 curren kernel: [<ffffffff81072fd1>] ? T.800+0x15/0x3b
Jan 14 06:40:59 curren kernel: [<ffffffff81074011>] ? ktime_get_ts+0x44/0x93
Jan 14 06:40:59 curren kernel: [<ffffffff81112067>] ? poll_select_copy_remaining+0xd2/0xf6
Jan 14 06:40:59 curren kernel: [<ffffffff811122d1>] ? sys_select+0xa4/0xb9
Jan 14 06:40:59 curren kernel: [<ffffffff816efb65>] page_fault+0x25/0x30
Jan 14 06:40:59 curren kernel: Mem-Info:
Jan 14 06:40:59 curren kernel: Node 0 DMA per-cpu:
Jan 14 06:40:59 curren kernel: CPU    0: hi:    0, btch:   1 usd:   0

here is result of free -m`

root@curren:/var/log# free -m
             total       used       free     shared    buffers     cached
Mem:           486        481          5          0          0          7
-/+ buffers/cache:        473         13
Swap:          511        507          4
root@curren:/var/log#

You really have run out of memory (and swap!).

Either tune your services to use less memory, or buy a VPS with more memory.

Check more discussion of this question.

I’m running nginx 1.2.6 behind amazon elb.

Periodically I see the following lines in nginx access.log:

10.28.135.252 - - [14/Jan/2013:14:20:07 +0000] "-" 400 0 "-" "-" "-"
10.28.135.252 - - [14/Jan/2013:14:20:07 +0000] "-" 400 0 "-" "-" "-"
10.28.135.252 - - [14/Jan/2013:14:20:14 +0000] "-" 400 0 "-" "-" "-"
10.28.135.252 - - [14/Jan/2013:14:20:14 +0000] "-" 400 0 "-" "-" "-"
10.28.135.252 - - [14/Jan/2013:14:20:18 +0000] "-" 400 0 "-" "-" "-"
10.28.135.252 - - [14/Jan/2013:14:20:48 +0000] "-" 400 0 "-" "-" "-"

There are no errors in the error log.
What can it be?

Thanks.

This occurs when a client connects to the web server, but then fails to send a request before disconnecting. It also occurs when a client connects, sends a request with Connection: Keep-Alive set, but disconnects before sending another request. In either case, they can be safely ignored.

Check more discussion of this question.

I use fedora 17, and when I setup nginx with uwsgi using unix domain socket, when I place the socket in a directory with proper permission it’s ok, but when I place the socket in the /tmp it will cause nginx error:

connect() to unix:/tmp/MySite.sock failed (2: No such file or directory) while connecting to upstream

The file does exist and has read/write permission for nginx user. But what cause this error, its really drive me crazy, can somebody figure it out.

You can’t place sockets intended for interprocess communication in /tmp.

For security reasons, recent versions of Fedora use namespaced temporary directories, meaning every service sees a completely different /tmp and can only see its own files in that directory.

To resolve the issue, place the socket in a different directory, such as /run (formerly known as /var/run).

Check more discussion of this question.