Godfatherhost

As you may know you can login to your SSH server with use of DSA/RSA keys without entering password. The simplest way to add key to the remote server is ssh-copy-id script. Here is an example of how you can add key if you use non standard SSH port:

ssh-copy-id -i /home/myuser/.ssh/id_dsa.pub “[email protected] -p 2222″

If you need to get list of available packages for update in Fedora you can use next command:

yum clean all && yum check-update

And the same for Debian:

apt-get clean all && apt-get -qqs dist-upgrade

RRDtool is a great facility which aims to replace MRTG and was written by Tobias Oetiker. RRDtool provides powerful features for collecting and visualizing various system metrics like network traffic, MySQL counters//http://godfatherhost.com/cloud/index.html or whatever you want. It’s always good idea to know what is going on under the hood of your server. Managing servers we prefer to monitor its parameters. Here is a short introduction into using RRD with Python. We will use Fedora 15  here but all examples should work in any another distributions as well. If you’re not familiar with RRD  you might need to check its tutorial.

At the beginning  we need to install RRD library for using with Python:

yum install rrdtool-python

Now we can create RRD database with all sources we need. Please note that here we use 30 minutes intervals, but you can use any you want. We use two parameters as data source: metric1 and metric2. It’s type is GAUGE. The thing is RRD supports different types (actually, Data Source Type): GAUGE, DERIVE, ABSOLUTE, COMPUTE or COUNTER. Probably, the most popular are COUNTER and GAUGE. The first one should be used for something  like SNMP network interface counters for input and output traffic which increase constantly. And when you use just some parameters like HDD or CPU temperature//http://godfatherhost.com/dedicated/index.htm you need to use GAUGE and RRD will gather data as is with no deltas. You can find other types description in RRD documentation.

01#!/usr/bin/python

02import rrdtool

03ret = rrdtool.create("example.rrd", "--step", "1800", "--start", '0',

04 "DS:metric1:GAUGE:2000:U:U",

05 "DS:metric2:GAUGE:2000:U:U",

06 "RRA:AVERAGE:0.5:1:600",

07 "RRA:AVERAGE:0.5:6:700",

08 "RRA:AVERAGE:0.5:24:775",

09 "RRA:AVERAGE:0.5:288:797",

10 "RRA:MAX:0.5:1:600",

11 "RRA:MAX:0.5:6:700",

12 "RRA:MAX:0.5:24:775",

13 "RRA:MAX:0.5:444:797")

Let’s consider all lines in details. First line include name of RRD database (“example.rrd”) and you can use here any path you want, step of parameters checking (30 minutes in our case), and the start point (0 or N means ‘now’).  ’DS’ in line 4-5 means Data Source, these lines include two our metrics. ’2000′ means that RRD can wait for 2000 seconds to get new values until it considers them as unknown (that’s is why we use 2000, which 200 seconds more of our 30 minutes interval).  Last two parameters – ‘U:U’ – stand for min and max values of each metric (‘unknown’ in our case). Lines 6-13 describe what types of gained values RRD should store in its database.  It’s pretty self-describing (average and max values).  Mentioned values describe how many parameters RRD should keep. Considering it can be confusing I will omit explanation but note that these values were choosen  to be compatible with MRTG (actually, it’s not quite true  since we use 1800 seconds periods and not 5 minutes, so you might need to modify it (if you also don’t use 5 minutes period) or keep like I did).

To update your RRD database use next example:

1from rrdtool import update as rrd_update

2ret = rrd_update('example.rrd', 'N:%s:%s' %(metric1, metric2));

‘N’ also means ‘now’. It should be launched every 30 minutes by cron, for instance. After some time RRD is ready to graph our data for a day, a week and a  month.

01import rrdtool

02for sched in ['daily' , 'weekly', 'monthly']:

03 

04    if sched == 'weekly':

05        period = 'w'

06    elif sched == 'daily':

07        period = 'd'

08    elif sched == 'monthly':

09        period = 'm'

10    ret = rrdtool.graph( "/var/www/html/metrics-%s.png" %(sched), "--start","-1%s" %(period), "--vertical-label=Num",

11         '--watermark=playground.in.supportex.net',

12         "-w 800",

13         "DEF:m1_num=example:metric1:AVERAGE",

14         "DEF:m2_num=example.rrd:metric2:AVERAGE",

15         "LINE1:m1_num#0000FF:metric1\\r",

16         "LINE2:m2_num#00FF00:metric2\\r",

17         "GPRINT:m1_num:AVERAGE:Avg m1\: %6.0lf ",

18         "GPRINT:m1_num:MAX:Max m1\: %6.0lf \\r",

19         "GPRINT:m2_num:AVERAGE:Avg m2\: %6.0lf ",

20         "GPRINT:m2_num:MAX:Max m2\: %6.0lf \\r")

You should get something like:

Here is a simple way to check if your database on master and slave server have the same data. We will usemk-table-checksum from maatkit tools for this purpose.

Note! Never do it on the production server! Or use it on your own risk! As stated:

mk-table-checksum executes queries that cause the MySQL server to checksum its data. This can cause significant server load. It is read-only unless you use the --replicate option, in which case it inserts a small amount of data into the specified table.

This trick works only for statement-based replication and with big databases can lead to significant overhead. To install mk-table-checksum run:

wget http://www.maatkit.org/trunk/mk-table-checksum

chmod u+x ./mk-table-checksum

./mk-table-checksum -uroot -pstong_password --create-replicate --replicate YOUR_DB.checksums --ignore-databases information_schema,mysql --replicate-database=YOUR_DB --empty-replicate-table --chunk-size=1M nocheck-replication-filters master.example.com

This commands will execute CHECKSUM TABLE statement on all tables and will create table YOUR_DB.checksums to store sums. Due to the replication the same statements will be processed on the slave and it will calculate its own versions of checksums. After that mk-table-checksum can perform some query which will be used to define if databases are the same. More information on that can be found atmaatkit documentation.

Now we are ready to do last check (at this time ’--replicate-check 1′ are being used) :

./mk-table-checksum -uroot -pstrong_password --create-replicate --replicate YOUR_DB.checksums --ignore-databases information_schema,mysql --replicate-database=YOUR_DB --empty-replicate-table --chunk-size=1M --nocheck-replication-filters --replicate-check 1  master.example.com

If there’s no output it means there’s no differences between the master and slave.

Ruby has gem, Python has pip and Lua has luarocks. Unfortunately, in Fedora 15 there is no rpm package of luarocks, so if you need it and don’t want to install it from sources, you will have to build it. Fortunately, there is a spec file written by Duboucher Thomas, so we don’t need to write it from scratch. There is also package review request for Fedora 12, but bugs weren’t solved still so you might not want use it on production servers. At the moment last version of luarocks is 2.0.5. Download luarocks sources to you SOURCES directory:

wget http://luarocks.org/releases/luarocks-2.0.5.tar.gz

Change luarocks version in spec file to “2.0.5″:

Version: 2.0.5

Build:

rpmbuild -bb ~/rpmbuild/SPECS/luarocks-1.0-1.spec

Install:

rpm -ivh ~/rpmbuild/RPMS/x86_64/luarocks-2.0.5-1.fc15.x86_64.rpm

And use:

[root@playground ~]# luarocks search geo

Search results: ===============

Rockspecs and source rocks: —————————

lua-geoip 0.1-1 (rockspec) – http://luarocks.org/repositories/rocks 0.1-1 (src) – http://luarocks.org/repositories/rocks

[root@playground ~]#

Lua is a tiny but powerful programming language. Due to its small memory footprint (about hundreds of kilobytes) it’s widely used in application which need to be extended with more complex logic. For instance, Salvatore Sanfilippo has published a  post describing how you can use power of Lua with Redis. It also applied to nginx, you can use Lua in its configuration files to get them more flexible. Sometimes it can be handy. Here is what you need to do to get nginx with Lua in core. We will use Lua module for nginx. At the end we will have nginx rpm package with Lua support. I assume that you already have rpmbuild environment and use Fedora 15. You also need to have Lua installed. Firstly download and unpack ngx_lua and to your SOURCES. You will also need ngx_devel_kit, it should be unpacked to SOURCES as well. After that you need to download srpm of nginx and install it.

yumdownloader –source nginx rpm -ivh nginx-1.0.5-1.fc15.src.rpm

Besides you need to download nginx tarboll with 1.0.6 version and put it  to SOURCES. Change Version line to “1.0.6″ and add next line to the configure section:

–add-module=/root/rpmbuild/SOURCES/simpl-ngx_devel_kit-bc97eea/ \ –add-module=/root/rpmbuild/SOURCES/chaoslawful-lua-nginx-module-752be5d/ \

Now we are ready to build rpm.

rpmbuild -bb ~/rpmbuild/SPECS/nginx.spec

If there are  no errors you should get nginx rpm package with Lua support:

[root@playground ~]# ls ~/rpmbuild/RPMS/x86_64/nginx-1.0.6-1.fc15.x86_64.rpm /root/rpmbuild/RPMS/x86_64/nginx-1.0.6-1.fc15.x86_64.rpm [root@playground ~]#

Let’s install it and test.

rpm -ivh ~/rpmbuild/RPMS/x86_64/nginx-1.0.6-1.fc15.x86_64.rpm

As you can see nginx was built with Lua:

[root@playground ~]# ldd /usr/sbin/nginx | grep lua liblua-5.1.so => /usr/lib64/liblua-5.1.so (0×0000003331600000) [root@playground ~]#

Add next section to nginx configuration file (thats is a slightly modified example from the documentation):

location /foo { set $diff ”; # we have to predefine the $diff variable here

set_by_lua $sum ‘ local a = 32 local b = 56

ngx.var.diff = a – b; — write to $diff directly return a + b; — return the $sum value normally ‘; ngx.print(ngx.var.sum); ngx.say(ngx.var.diff);’; }

Start nginx and make a request:

[root@playground ~]# /etc/init.d/nginx start ; curl localhost/foo Restarting nginx (via systemctl): [ OK ] 88-24 [root@playground ~]#

Usually to set up MySQL replication it’s convenient to dump database from current master with –master-dataparameter. With this option mysqldump includes commented SQL operator which needs to be issued on slave to set up replication:

CHANGE MASTER TO MASTER_LOG_FILE=’mysqld-bin.000008′, MASTER_LOG_POS=687808977;

But sometime it’s not very good idea to dump database from master server. For instance, you use MyISAM tables and you don’t want to lock them while dumping. In this case you can dump database from on e of the slaves (of course, if you already have them). That is what you need to do. Stop SQL_THREAD:

STOP SLAVE SQL_THREAD;

Check slave status:

SHOW SLAVE STATUS\G;

Get master binary log coordinates:

mysql> pager egrep ‘(Relay_Master_Log_File|Exec_Master_Log_Pos)’ ; show slave status\G; PAGER set to ‘egrep ‘(Relay_Master_Log_File|Exec_Master_Log_Pos)” Relay_Master_Log_File: mysql-bin.000225 Exec_Master_Log_Pos: 632106728 1 row in set (0.00 sec)

ERROR: No query specified

mysql>

Save last output to use it later. At thus point we can start slave:

START SLAVE SQL_THREAD;

Make sure it’s working:

SHOW SLAVE STATUS\G;

After restoring database on new slave change its master to your current one using values above:

CHANGE MASTER TO MASTER_LOG_FILE=’mysql-bin.000225′,MASTER_LOG_POS=632106728;

First value is Relay_Master_Log_File, and the second one is Exec_Master_Log_Pos.

In most case simple master-slave replication provides enough level of high-avalability (about 99.9%), read ops balancing and allows non-blocking backups (if you use MyISAM tables since InnoDB tables can be//http://godfatherhost.com/cloud/index.html backed up without blocking). But sometimes you may want to setup chained replication with two slaves: first slave replicates from the master, and the second slave replicates from the first one. Often such technique is used for smooth MySQL server migration. In this this case first slave is switched to master mode, and second slave remains the slave. Hereby we will refer master as M1, first slave as S1, and the second one as S2, so S1 replicates from M1, and S2 replicates from S1.

M1

Enable binary logging:

server-id=1 log-bin=mysql-bin binlog_do_db=sakila

Provide privileges for slave:

GRANT REPLICATION SLAVE ON *.* TO ‘replica_user’@'slave1.example.com’ IDENTIFIED BY ‘password’; FLUSH PRIVILEGES;

Dump database:

mysqldump -uroot –opt –master-data -p sakila > /tmp/sakila.sql

M1 is ready. Going to S1.

S1

server-id=2 log-bin=mysql-bin replicate-do-db=sakila log-slave-updates skip-slave-start

Missing log-slave-updates parameter is the most common mistake while setting up chained replication. So make sure you have it. Granting privileges:

GRANT REPLICATION SLAVE ON *.* TO ‘replica_user’@'slave2.example.com’ IDENTIFIED BY ‘password’; FLUSH PRIVILEGES;

Drop replicated database sakila if exists:

DROP DATABASE sakila;

Create empty database:

CREATE DATABASE sakila;

Reset old slave:

RESET SLAVE;

Import dump from M1 to S1:

mysql -uroot -p sakila < /tmp/sakila.sql

Obtaining M1 binary log coordinates. Thet can be found at the bginning of dump file:

CHANGE MASTER TO MASTER_LOG_FILE=’mysqld-bin.000002′, MASTER_LOG_POS=107;

Providing information about replication user:

CHANGE MASTER TO MASTER_HOST=’master1.example.com’, MASTER_USER=’replica_user’, MASTER_PASSWORD=’password’;

Checking slave status:

SHOW SLAVE STATUS\G

Let’s continue with S2.

S2

Adding replication parameters to S2:

server-id=3 log-bin=mysql-bin binlog_do_db=sakila replicate-do-db=sakila

Drop database sakila if exists:

DROP DATABASE sakila;

Create empty database:

CREATE DATABASE sakila;

Reset old slave:

RESET SLAVE;

Import dump from M1 to S2:

mysql -uroot -p sakila < /tmp/sakila.sql

At this moment we need to obtain S1 binary log coordinates in order to start replicating from it.

Run ‘SHOW MASTER STATUS’ on S1 to get them:

mysql> SHOW MASTER STATUS; +------------------+-----------+--------------+------------------+ | File| Position | Binlog_Do_DB | Binlog_Ignore_DB | +------------------+-----------+--------------+------------------+ | mysql-bin.000033 | 743915402 | | | +------------------+-----------+--------------+------------------+ 1 row in set (0.00 sec)

mysql>

Provide master information to S2:

CHANGE MASTER TO MASTER_HOST=’slave1.example.com’, MASTER_USER=’replica_user’, MASTER_PASSWORD=’password’;

Providing S1 binary log infromation to S2:

CHANGE MASTER TO MASTER_LOG_FILE=’mysql-bin.000033′, MASTER_LOG_POS=743915402;

Starting replication on S2:

START SLAVE; SHOW SLAVE STATUS\G;

mysql> SHOW SLAVE STATUS\G; *************************** 1. row *************************** Slave_IO_State: Waiting for master to send event Master_Host: slave1.example.com Master_User: replica_user Master_Port: 3306 Connect_Retry: 60 Master_Log_File: mysql-bin.000033 Read_Master_Log_Pos: 747231154 Relay_Log_File: relay-bin.000097 Relay_Log_Pos: 747231299 Relay_Master_Log_File: mysql-bin.000033 Slave_IO_Running: Yes Slave_SQL_Running: Yes Replicate_Do_DB: sakila Replicate_Ignore_DB: Replicate_Do_Table: Replicate_Ignore_Table: Replicate_Wild_Do_Table: Replicate_Wild_Ignore_Table: Last_Errno: 0 Last_Error: Skip_Counter: 0 Exec_Master_Log_Pos: 747231154 Relay_Log_Space: 747231494 Until_Condition: None Until_Log_File: Until_Log_Pos: 0 Master_SSL_Allowed: No Master_SSL_CA_File: Master_SSL_CA_Path: Master_SSL_Cert: Master_SSL_Cipher: Master_SSL_Key: Seconds_Behind_Master: 0 Master_SSL_Verify_Server_Cert: No Last_IO_Errno: 0 Last_IO_Error: Last_SQL_Errno: 0 Last_SQL_Error:

Now we can start slave threads on S1 as well:

START SLAVE; SHOW SLAVE STATUS\G;

What to monitor with replication

Installing OTRS on Fedora is quite easy considering it has pre-built RPM packages and clean//http://godfatherhost.com/cloud/index.html documentation.

Usually OTRS is used with Apache. It’s stable well-tested solution. But what to do if you have nginx and you don’t want to have Apache? An issue becomes more complicated considering that nginx doesn’t support CGI from a box, and OTRS can’t be deployed as FastCGI application. If you try to use OTRS in FastCGi mode you will get such error messages:

Wide character in FCGI::Stream::PRINT at /opt/otrs/bin/fcgi-bin/../../Kernel/Output/HTML/Layout.pm line 1663. Wide character in FCGI::Stream::PRINT at /opt/otrs/bin/fcgi-bin/../../Kernel/Output/HTML/Layout.pm line 1663″

And it caused by CGI::Fast Perl library newer than 0.67. Anyway, it’s not a good idea to downgrade it. Fortunately, we can use FastCGI wrapper, so OTRS will be FastCGI application for nginx, but will be launched as CGI.

There are two FastCGI wrappers for using CGI application with nginx. First one can be found on nginx wiki page.  Also you can use modified by Denis S. Filimonov version of this wrapper. It’s written in Pelr and can be found here. The second wrapper is fcgiwrap. It’s writen in C. I have choosen Perl wrapper since  there is no need to compile sources.

Put wrapper to /usr/local/bin/fastcgi-wrapper.pl and change it to open Unix socket at /var/run/otrs/perl_cgi-dispatch.sock:

$socket = FCGI::OpenSocket( “/var/run/otrs/perl_cgi-dispatch.sock”, 10 );

Here is systemd configuration file for fast-cgi wrapper:

[Unit] Description=otrs-index.pl

[Service] ExecStart=/usr/local/bin/fastcgi-wrapper.pl Type=forking User=otrs Group=nginx Restart=always

[Install] WantedBy=multi-user.target

Put it to /lib/systemd/system/perl-fcgi.service. Now we need to enable it:

systemctl enable  perl-fcgi.service

nginx configuration file:

server { listen 1.2.3.4:80; server_name otrs.example.com otrs; root /opt/otrs/var/httpd/htdocs; index index.html;

location /otrs-web { gzip on; alias /opt/otrs/var/httpd/htdocs; }

location ~ ^/otrs/(.*\.pl)(/.*)?$ { fastcgi_pass unix:/var/run/otrs/perl_cgi-dispatch.sock; fastcgi_index index.pl; fastcgi_param SCRIPT_FILENAME /opt/otrs/bin/fcgi-bin/$1; fastcgi_param QUERY_STRING $query_string; fastcgi_param REQUEST_METHOD $request_method; fastcgi_param CONTENT_TYPE $content_type; fastcgi_param CONTENT_LENGTH $content_length; fastcgi_param GATEWAY_INTERFACE CGI/1.1; fastcgi_param SERVER_SOFTWARE nginx; fastcgi_param SCRIPT_NAME $fastcgi_script_name; fastcgi_param REQUEST_URI $request_uri; fastcgi_param DOCUMENT_URI $document_uri; fastcgi_param DOCUMENT_ROOT $document_root; fastcgi_param SERVER_PROTOCOL $server_protocol; fastcgi_param REMOTE_ADDR $remote_addr; fastcgi_param REMOTE_PORT $remote_port; fastcgi_param SERVER_ADDR $server_addr; fastcgi_param SERVER_PORT $server_port; fastcgi_param SERVER_NAME $server_name; } }

Restating nginx, startting FastCGI wrapper:

service nginx restart service perl-fcgi start

Nagios is powerful monitoring software. And like OTRS in most cases it’s used with Apache. But//http://godfatherhost.com/cloud/index.html sometimes you might want to use it with nginx. Here is a simple way to do that. Actually it’s very similar to launching OTRS with nginx – in both ways we use Perl FastCGI wrapper. I assume that you already have Nagios installed and use Fedora 15.  Download Perl FastCGI wrapper and put it to /usr/local/bin/fastcgi-wrapper.pl.  Modify it to use TCP socket:

$socket = FCGI::OpenSocket( “127.0.0.1:9002″, 10 ); #use IP sockets

Put this systemd configuration file to /lib/systemd/system/fcgi-perl.service:

[Unit] Description=fastcgi-wrapper.pl [Service] ExecStart=/usr/local/bin/fastcgi-wrapper.pl Type=forking User=apache Group=apache Restart=always [Install] WantedBy=multi-user.target

Enable it:

systemctl enable fcgi-perl.service

For using Nagios web interface you also need to have PHP processes pool, in our case it listens TCP socket 127.0.0.1:9005. So if you don’t have one, you need to launch it (convenient way to do that is to use spawn-fcgi tool).

Add next lines to your nginx server configuration file:

location /nagios { alias /usr/share/nagios/html; index index.php; }

location ~ ^/nagios/(.*\.php)$ { root /usr/share/nagios/html/; rewrite ^/nagios/(.*) /$1 break; fastcgi_index index.php; include /etc/nginx/fastcgi_params; fastcgi_param SCRIPT_FILENAME /usr/share/nagios/html$fastcgi_script_name; fastcgi_pass 127.0.0.1:9005; }

location ~ ^/nagios/cgi-bin/ { root /usr/lib64/nagios/cgi-bin/; rewrite ^/nagios/cgi-bin/(.*)\.cgi /$1.cgi break; include /etc/nginx/fastcgi_params; fastcgi_param AUTH_USER $remote_user; fastcgi_param REMOTE_USER $remote_user; fastcgi_param SCRIPT_FILENAME /usr/lib64/nagios/cgi-bin$fastcgi_script_name; fastcgi_pass 127.0.0.1:9002;

}

service nginx restart

To get statistics//http://godfatherhost.com/cloud/index.html information about query numbers that Bind has served you need to add next line to configuration file:

statistics-file “/var/stats/named.stats”;

It should be added in a ‘global’ options statement. After that you need to restart Bind and run:

rndc stats

Bind will dump statistics to this file. Note that if you// http://godfatherhost.com/dedicated/index.html run Bind chrooted it will be in a chroot filesystem.

[root@playground ~]# cat /var/named/var/stats/named.stats

From the box nginx doesn’t support CGI applications deployment since CGI is very old and poor//http://godfatherhost.com/dedicated/index.html performance way to run them. But some applications still should be deployed with use of CGI (earlier, we covered setup ofNagios and OTRS with nginx). Here is another way to do that. In this case we will use fcgiwrap tool by Grzegorz Nosek. It’s written in C and should provide better performance in comparison with pure Perl wrapper we use before.

Firstly we need install it. If you use Debian or Ubuntu just install if with apt-get:

aptitude install fcgiwrap

For Fedora we’ll need compile it from sources. Download fcgiwrap from Github and unpack. Install fcgi-devel:

yum install fcgi-devel

Then run:

make install

Now you should have fcgiwrap installed in /usr/local/bin/. So, we can proceed with nginx. Here is a sample configuration. As you can see all .cgi files will be considered as CGI applications.

server { listen 80; server_name example.com; root /home/web; index index.html index.cgi; error_log /var/log/nginx/error.log; access_log /var/log/nginx/access.log main;

location / { root /home/web; index index.html index.htm; }

location ~ ^/cgi-bin/.*\.cgi$ { fastcgi_pass unix:/var/run/nginx/cgiwrap-dispatch.sock; fastcgi_index index.cgi; fastcgi_param SCRIPT_FILENAME /home/web/$fastcgi_script_name; fastcgi_param QUERY_STRING $query_string; fastcgi_param REQUEST_METHOD $request_method; fastcgi_param CONTENT_TYPE $content_type; fastcgi_param CONTENT_LENGTH $content_length; fastcgi_param GATEWAY_INTERFACE CGI/1.1; fastcgi_param SERVER_SOFTWARE nginx; fastcgi_param SCRIPT_NAME $fastcgi_script_name; fastcgi_param REQUEST_URI $request_uri; fastcgi_param DOCUMENT_URI $document_uri; fastcgi_param DOCUMENT_ROOT $document_root; fastcgi_param SERVER_PROTOCOL $server_protocol; fastcgi_param REMOTE_ADDR $remote_addr; fastcgi_param REMOTE_PORT $remote_port; fastcgi_param SERVER_ADDR $server_addr; fastcgi_param SERVER_PORT $server_port; fastcgi_param SERVER_NAME $server_name; } }

At this step we need some way to spawn our FCGI wrapper instances. For that we will use spawn-fcgi:

spawn-fcgi -u web -g web -M 0775 -F 10 -s /var/run/nginx/cgiwrap-dispatch.sock -U web -G web /usr/local/bin/fcgiwrap

To use it we will also need some kind of init.d script. Modern versions of Fedora ship with systemd, so we can use it for managing of FCGI instances pool:

[Unit] Description=spawn-perl-fcgi

[Service] ExecStart=/usr/bin/spawn-fcgi -u web -g web -M 0775 -F 10 -s /var/run/nginx/cgiwrap-dispatch.sock -U web -G web /usr/local/bin/fcgiwrap Type=forking User=root Group=root Restart=always

[Install] WantedBy=multi-user.target

Put in in /lib/systemd/system/spawn-perl-fcgi.service and run:

systemctl enable spawn-perl-fcgi.service

For other systems you can use script provided in nginx wiki.

This is it.

If you have several MySQL servers it makes sense to use one phpMyAmin to manage them.  To do it just add as many server/http://godfatherhost.com/cloud/index.html/ as you need to config.inc.php. In Fedora and Centos it is located in /etc/phpMyAdmin/. Here’s an example:

<?php

/* Servers configuration */ $i = 0;

/* Server: localhost [1] */ $i++; $cfg['Servers'][$i]['verbose'] = ”; $cfg['Servers'][$i]['host'] = 10.20.90.1; $cfg['Servers'][$i]['port'] = ”; $cfg['Servers'][$i]['socket'] = ”; $cfg['Servers'][$i]['connect_type'] = ‘tcp’; $cfg['Servers'][$i]['extension'] = ‘mysqli’; $cfg['Servers'][$i]['auth_type'] = ‘cookie’;

/* Server: 10.20.90.2 [2] */ $i++; $cfg['Servers'][$i]['verbose'] = ”; $cfg['Servers'][$i]['host'] = ’10.20.90.2′; $cfg['Servers'][$i]['port'] = ”; $cfg['Servers'][$i]['socket'] = ”; $cfg['Servers'][$i]['connect_type'] = ‘tcp’; $cfg['Servers'][$i]['extension'] = ‘mysqli’; $cfg['Servers'][$i]['auth_type'] = ‘cookie’;

/* End of servers configuration */

$cfg['blowfish_secret'] = ’3242fndsnj342qljds32s’; $cfg['DefaultLang'] = ‘en’; $cfg['ServerDefault'] = 0; $cfg['UploadDir'] = ”; $cfg['SaveDir'] = ”; ?>

Now you can select which MySQL server to use before you logged in. For security reasons it’s always better to use only SSL hosts, non-root MySQL accounts and limit access to phpMyAdmin at least with IP addresses.

Often there’s a need to get changelog of RPM//http://godfatherhost.com/cloud/index.html package. Here is a way to do that:

[root@playground ~]# rpm -qp –changelog kernel-2.6.41.9-1.fc15.x86_64.rpm | head * Fri Jan 13 2012 Josh Boyer <[email protected]> 2.6.41.9-1 - Linux 3.1.9 - CVE-2012-0045 kvm: syscall instruction induced guest panic (rhbz 773392)

* Wed Jan 11 2012 Josh Boyer <[email protected]> - Patch from Stanislaw Gruszka to fix NULL ptr deref in mac80211 (rhbz 769766)

* Tue Jan 10 2012 Josh Boyer <[email protected]> - Fix ext3/ext4 corrupted bitmap error path (pointed out by Eric Sandeen)

[root@playground ~]#

If you get an error

ERROR 1005 (HY000) at line 14: Can’t create table example.tbl (errno: 150)

it could be for at least two reasons. MySQL doesn’t allow//http://godfatherhost.com/cloud/index.html to create foreign keys for  a set of tables one of which doesn’t exist. To solve this restriction you can set SET foreign_key_checks to 0:

mysql> SET foreign_key_checks = 0;

And after you done with restoring you should enable it back:

mysql> SET foreign_key_checks = 1;

The second reason could be different data types //http://godfatherhost.com/dedicated/index.htmlof columns you trying to create constrain between. For instance, bigint(20) and bigint(21). So, keep it in mind.

More information on  FOREIGN KEY Constraints.

If you get mentioned error while//http://godfatherhost.com/dedicated/index.html restoring big database, make sure you have enough size ofmax_allowed_packet.

[mysqld] max_allowed_packet=64M

Let’s consider typical LEMP stack:

Linux

nginx

php-fpm

MySQL server

For security reasons you might want to disable some dangerous functions like exec or system.  As you may know, you// http://godfatherhost.com/dedicated/index.htmlcan do this by adding list of function to disable_functions parameter in your php.ini. At the same time php-fpm allows to limit PHP parameters per pool. So, we go to our pool config (for instance, /etc/php-fpm.d/production.conf) and try to add:

php_admin_value[disable_functions] = dl,exec,passthru,shell_exec,system,proc_open,popen,curl_exec,curl_multi_exec,parse_ini_file,show_source

And here you can find out that//http://godfatherhost.com/cloud/index.html it doesn’t work if you have suhosin installed. To solve this issue use suhosin.executor.func.blacklist in suhosin config file (which should be in /etc/php.d/suhosin.ini on Fedora/Centos).

suhosin.executor.func.blacklist = “dl,exec,passthru,shell_exec,system,proc_open,popen,curl_exec,curl_multi_exec,parse_ini_file,show_source”