Difference between revisions of "Service/backup"

From Sugar Labs
Jump to navigation Jump to search
 
(21 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
== Locations ==
 
== Locations ==
  
* /backup on treehouse.sugarlabs.org (for sunjammer)
+
* /backup on [[Machine/papert]]: daily filesystem backups for [[Machine/bender]] and [[Machine/sunjammer]]
* backup.sugarlabs.org (all the other systems)
+
* /backup on [[Machine/bender]]: daily filesystem backups for all VMs hosted on [[Machine/papert]]
* /backup on sunjammer.sugarlabs.org (of sunjammers services: git, mysql and pgsql)
 
 
 
== backup.sugarlabs.org ==
 
 
 
[[Machine/owl
 
owl.laptop.org:/home/sugarbackup/ ]]
 
 
 
  
 
== Administrative contact ==
 
== Administrative contact ==
 
bernie AT sugarlabs DOT org
 
 
 
== Sysadmins ==
 
 
For non-emergency calls, preferably send email to the administrative contact.
 
 
 
* [[User:Bernie|Bernie Innocenti]]
 
* [[User:Bernie|Bernie Innocenti]]
  
 +
== wizbackup ==
  
== Restore Backups ==
+
* [https://codewiz.org/gitweb/wizbackup.git Source code]
 
+
* [https://codewiz.org/gitweb/wizbackup.git/blob/HEAD:/README.md README]
To restore a backup try something like:
 
 
 
  duplicity restore -t 7D --restore-time <yyyy/mm/dd --file-to-restore <source/path/to/restore/from>  scp://sugarbackup@backup.sugarlabs.org/backup/<machine.sugarlabs.org>  </destination/path/to/restore/to>
 
 
 
* Replace <> with your values
 
 
 
 
 
Example:
 
 
 
  duplicity restore -t 7D --restore-time 2000/01/01 --file-to-restore var/lib/mysql/wordpress  scp://sugarbackup@backup.sugarlabs.org/backup/sunjammer.sugarlabs.org  /home/bernie/restore/wordpress
 
 
 
 
 
the --restore-time is a workaround for a bug in duplicity when restoring backups in the past
 
 
 
 
 
 
 
== Notes ==
 
 
 
This is the main backup server for all [[Machine|Sugarlabs Machines]]
 
 
 
  
== Upgrade notes ==
+
This is the main backup script for all [[Machine|Sugar Labs Machines]]. It's a minimalist script based on rsync and keeps a history of N days using hard-links. While it's very very simple, wizbackup gets the job done and has the added benefit of not requiring any tools to read or restore a backup. The advantage over similar backup solution is that backups are just plain filesystem snapshots of the source machines; the main disadvantage is poor performance with large files which change often, such as logs or databases (my solution is to exclude these and backup the data using ad-hoc methods, such as SQL dumps).
  
 +
The `wizbackup-driver` script runs daily on the backup servers from `/etc/cron.daily/wizbackup`.
 +
It reads a list of hostnames to backup from `/backup/HOSTS/<machine-group-name>` and an optional list of files exclusions from `/backup/EXCLUDES`, in the same format taken by the `--exclude` flag of rsync.
  
 +
For each machine in the list, `wizbackup-driver` invokes `wizbackup` with this command-line:
  
== See Also ==
+
  wizbackup <hostname>:/ <local-backup-dir> --exclude-file /backup/EXCLUDE/ALWAYS --exclude-file /backup/EXCLUDE/<hostname>
  
* There are instructions on how to set up a backup here: http://wiki.sugarlabs.org/go/Infrastructure_Team/Template_virtual_machine_lucid
+
Wizbackup uses rsync and ssh to connect to the remote machine. Hence, the backup server must be authorized to connect as root to all backup hosts (e.g.: by storing the identity in `/root/.ssh/authorized_keys`). Wizbackup reads the ssh private key from `/etc/wizbackup/ssh_id`).
  
 
[[Category:Service|zzz]]
 
[[Category:Service|zzz]]

Latest revision as of 15:09, 2 July 2024

Locations

Administrative contact

wizbackup

This is the main backup script for all Sugar Labs Machines. It's a minimalist script based on rsync and keeps a history of N days using hard-links. While it's very very simple, wizbackup gets the job done and has the added benefit of not requiring any tools to read or restore a backup. The advantage over similar backup solution is that backups are just plain filesystem snapshots of the source machines; the main disadvantage is poor performance with large files which change often, such as logs or databases (my solution is to exclude these and backup the data using ad-hoc methods, such as SQL dumps).

The `wizbackup-driver` script runs daily on the backup servers from `/etc/cron.daily/wizbackup`. It reads a list of hostnames to backup from `/backup/HOSTS/<machine-group-name>` and an optional list of files exclusions from `/backup/EXCLUDES`, in the same format taken by the `--exclude` flag of rsync.

For each machine in the list, `wizbackup-driver` invokes `wizbackup` with this command-line:

 wizbackup <hostname>:/ <local-backup-dir> --exclude-file /backup/EXCLUDE/ALWAYS --exclude-file /backup/EXCLUDE/<hostname>

Wizbackup uses rsync and ssh to connect to the remote machine. Hence, the backup server must be authorized to connect as root to all backup hosts (e.g.: by storing the identity in `/root/.ssh/authorized_keys`). Wizbackup reads the ssh private key from `/etc/wizbackup/ssh_id`).