Script Webseite Duplizieren / Testseite: Unterschied zwischen den Versionen

Aus revampedia
Zeile 25: Zeile 25:
 
=== Script ===
 
=== Script ===
  
'''''WIP / in development / untested'''''
+
''WIP / in development / untested''
  
<code>./website-copy</code>:
+
Code on github: https://github.com/rebootl/website-copy/
<source lang="bash">
 
#!/bin/bash
 
#
 
# script to copy websites incl. database (mysql for now)
 
# from remote to local server
 
#
 
# e.g. from production to test servers / creating a test-site,
 
# might later also be used as a base for a failover sync script
 
#
 
# usage (on local server);
 
#
 
#  ./website-copy <config-file>
 
#
 
# using one config file per-website
 
# multiple copies of one website can be created under different
 
# names / using separate configs
 
#
 
# required permissions:
 
#
 
# - ssh access is required on the remote host(s)
 
#  (incl. read permission for website files)
 
#  atm. the user calling the script is used
 
#
 
# - sudo local for writing website files with
 
#  configured user
 
#
 
# - mysql access must be configured using ~/.my.cnf files
 
#  on the local and the remote host
 
#  ~/.my.cnf format:
 
#  [mysqldump]
 
#  user=mysqluser
 
#  password=mysecretsqlpassword
 
#
 
# files are transferred using rsync (--> encrypted ?)
 
#
 
# the database is safely transferred using ssh
 
# (encrypted and compressed using gzip)
 
#
 
# using rsync, ssh, gzip
 
# mysqldump, mysql/mysqladmin (?)
 
#
 
# cem, 2016-09-01
 
#
 
#
 
# program workflow:
 
#
 
# 1. read settings from config file
 
#
 
# 2. copy files using rsync
 
#
 
# 3. copy database
 
#  3.1 create new db
 
#  3.2 copy from remote
 
#
 
#
 
#
 
# $1 config file
 
 
 
# check command line argument for config file
 
 
 
if [ $# -lt 1 ]; then
 
    echo "error: config file missing"
 
    echo "usage: $0 <config-file>"
 
    exit 1
 
fi
 
 
 
if [ ! -f "$1" ]; then
 
    echo "error: file not found: $1"
 
    exit 1
 
fi
 
 
 
# import settings
 
. "$1"
 
 
 
ME=$(whoami)
 
 
 
check_exit() {
 
    # $1: exit code, $?
 
    if [ $1 -ne 0 ]; then
 
        echo "oops, something went wrong, exiting."
 
        exit 1
 
    else
 
        echo "done."
 
    fi
 
}
 
 
 
# copy files from server using rsync
 
# rsync opts
 
# -r recursive
 
# -l copy symlinks as symlinks
 
# -p preserve permissions
 
#
 
# remote path includes always a trailing slash,
 
# by `parameter substitution`, e.g.: "${var%/}"/
 
# stripping if present and readding to avoid doubles
 
#
 
echo "copying files..."
 
sudo -u "$WEBSITE_LOCAL_USER". \
 
    rsync -rlpv "$ME"@"$WEBSITE_REMOTE_SERVER":"${WEBSITE_REMOTE_PATH%/}"/ \
 
    "$WEBSITE_LOCAL_PATH"
 
check_exit $?
 
 
 
# --> must create local db first ????
 
# create new db on local instance
 
# mysql > 5.7.6
 
# for mysql below this use something like:
 
# GRANT ALL ON `database`.* TO 'user'@'localhost' IDENTIFIED BY 'password';
 
# - http://stackoverflow.com/questions/13357760/mysql-create-user-if-not-exists#
 
echo "creating local database..."
 
mysql <<EOF
 
  CREATE DATABASE IF NOT EXISTS '$DB_LOCAL_NAME';
 
  CREATE USER IF NOT EXISTS '$DB_LOCAL_USER'@'localhost'
 
    IDENTIFIED BY '$DB_LOCAL_PASS';
 
  GRANT ALL PRIVILEGES ON $DB_LOCAL_NAME.* TO '$DB_LOCAL_USER'@'localhost';
 
  FLUSH PRIVILEGES;
 
EOF
 
check_exit $?
 
 
 
# ref: http://tltech.com/info/mysql-via-ssh/
 
# mysql destdb < <(ssh sourceserver.example.com "mysqldump sourcedb")
 
# sudo ??
 
# or ssh sourceserver.example.com "mysqldump sourcedb" | mysql destdb
 
#
 
# tested: ssh rochen "mysqldump joomla_revamp_ut| gzip" > joomla_revamp_ut.dump.gz
 
# works!
 
echo "copying database from remote..."
 
ssh "$DB_REMOTE_SERVER" "mysqldump $DB_NAME | gzip -c" \
 
    | gzip -d > mysql "$DB_LOCAL_NAME"
 
check_exit $?
 
 
 
# copying files from override directory
 
echo "copying files from override directory..."
 
sudo -u "$WEBSITE_LOCAL_USER". \
 
    cp -dpr "${OVERRIDE_PATH%/}"/. "$WEBSITE_LOCAL_PATH"
 
check_exit $?
 
 
 
echo "finished!"
 
 
 
</source>
 
 
 
<code>./example-config</code>:
 
<source lang="bash">
 
#!/bin/bash (only for syntax highlighting, should not be executable)
 
#
 
# per-website settings for website-copy script
 
# (one config per website)
 
#
 
# website: example
 
#
 
 
 
# per-website settings
 
 
 
# local settings
 
 
 
# local path of the website
 
# usually /var/www/something,
 
# website files will be written there using sudo -u with
 
# below local user (owner/group)
 
WEBSITE_LOCAL_PATH="/home/caydin/Scripte/website-copy/revamp_test"
 
WEBSITE_LOCAL_USER="www-data"
 
 
 
# local path of override files
 
# files and directories in this path are
 
# copied to the website local path,
 
# use this to overwrite config files etc.
 
OVERRIDE_PATH="/home/caydin/Scripte/website-copy/revamp_test-ovrd"
 
 
 
# local database information
 
# a new database will be created if it not already exists,
 
# and remote database tables are inserted here
 
# todo: --> HAS_DB=true/false
 
DB_LOCAL_NAME="joomla_revamp_test"
 
DB_LOCAL_USER="joomlauser"
 
DB_LOCAL_PASS="secret1.x"
 
 
 
# remote settings
 
 
 
# server and path of the website to copy
 
# rsync is used to log-in and copy the files from there
 
WEBSITE_REMOTE_SERVER="rochen.revamp-it.ch"
 
WEBSITE_REMOTE_PATH="/var/www/revamp/jodir_ut/"
 
 
 
# remote database information
 
# the database is copied encrypted and compressed
 
DB_SERVER="rochen.revamp-it.ch"
 
DB_NAME="joomla_revamp_ut"
 
 
 
</source>
 

Version vom 2. September 2016, 15:12 Uhr

Es soll ein Script erstellt werden um Webseiten inkl. Datenbank von produktiv- zu testserver zu kopieren, also eine testseite zu erstellen.

Geplantes Set-up:

Produktivserver
- webserver
- db-server

Testserver
- web- und db-server

Recherche

Google: "mysqldump over internet"

Script

WIP / in development / untested

Code on github: https://github.com/rebootl/website-copy/