Difference between revisions of "Setup MediaWiki on AWS"
| Line 418: | Line 418: | ||
</source> | </source> | ||
| + | =Setup S3cmd on hadoop0= | ||
| + | * install s3cmd | ||
| + | * s3cmd --configure | ||
| + | ::* Access Key: AKIAINNPVCF6X257DRPQ | ||
| + | ::* Secret Key: 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 | ||
| + | ::* encryption key: {eZDfswapkcoTXZoKBxLKw2ZmNNwiTKb | ||
| + | * works | ||
| + | * for first backup, go to hadoop0, in 1TB, create s3 dir, then backup-dft-hadoop02 subdir, cd there and issue command: | ||
| + | |||
| + | s3cmd sync -r s3://backup-dft-hadoop02 . | ||
| + | |||
| + | : and it backs s3 down. | ||
| + | * created crontab entry on hadoop0: | ||
| + | |||
| + | 0 5 * * 0 /usr/bin/s3cmd sync -q -r s3://backup-dft-hadoop02 /mnt/1TB/s3/backup-dft-hadoop02/ | ||
| + | |||
| + | |||
</onlydft> | </onlydft> | ||
<br /> | <br /> | ||
Revision as of 17:52, 22 January 2018
D. Thiebaut (talk) 16:44, 13 January 2018 (EST)
|
This page contains very sketchy steps for setting up (in my case moving) an existing mediawiki installation to AWS. This list is mostly for myself to remember what steps I have taken. |
|
Contents
Mediawiki Installation/Upgrade
- Follow the steps from https://www.mediawiki.org/wiki/Manual:Upgrading
- Download the new mediawiki in a directory belonging to the Web server
- unzip/untar.
- Create new MySql user/password/database on localhost (same host as Web server)
- Verify that base directory of mediawiki installation is accessible from a browser by putting a phpinfo.php file in it and loading it up with browser.
- use browser and point to /mw-config in the mediawiki web directory. Enter database information.
- Update the LocalSettings.php file
- If migrating, copy the extensions, skins, images, media, Downloads directory from old wiki to new installed wiki.
- Install one extension at a time in extensions directory, and at the end of LocalSettings.php
Updated Extensions
I had to update several extensions to match the new format.
AWS
EC2
- Go for free-tier t2-micro with attached EBS drive (20GB)
- Set OS to Ubuntu Server 16.04.
-
...
- Install MediaWiki on it, without data
- Install emacs
- Make an AMI out of it
- Move the images, extensions, skins, etc to newly installed MW.
- Install ddclient.
...
S3
- Create a bucket for backups
...
Backups
- Install duplicity and duply
- In the process create a GnuPG key for encryption of backups on S3
duply test create
- creates a profile to backup for MediaWiki in /data/html/dftwiki3
- All information in /home/ubuntu/.duply/test/conf
- Duplicity conf file:
- create script in /usr/local/bin, called backupDftWiki3MySql.sh to do mysqldump of MW to an sql file, and store it:
#! /bin/bash
# This will be called from a root cron job
mysqldump --user=thiebaut --password=xxxxxxxxxx dftwiki3 > /data/html/dftwiki3/backups/dftwiki3.sql 2> /dev/null
cd /data/html/dftwiki3/backups
gzip -f dftwiki3.sql
Route 53
- Created new domain dominiquefthiebaut.com
- Create record to map it to EC2 instance.
- Find public IP of EC2
...
and put in edit-box in record. Accept defaults.
- Disable ddclient on EC2 and with dyndns.
- Change the name of the host in /etc/hosts from hadoop02 to new hostname.
Crontab
Backup MediaWiki EC2 on AWS S3
- Crontab runs as root on ubuntu EC2
# ------------------------------------------------ # use mysqldump to dump mysql database for dftwiki3 # into a file called dftwiki3.sql in /data/html/dftwiki/backups 0 3 * * * /usr/local/bin/backupDftWiki3MySql.sh # ------------------------------------------------ # use duplicity (via duply) to backup /data/html/dftwiki3's main # directories to aws s3. The configuration and information about # which S3 bucket is in ~ubuntu/.duply/test/conf 0 4 * * * env HOME=/home/ubuntu duply test backup
Script to Copy MediaWiki from Xgridmac2 to EC2
Raw Log
