5 Responses to “Plesk Backups” »»

1. Nick
   Comment by Nick | 06/02/07 at 11:23 pm

   Many thanks - very useful to know for a linux/plesk newbie like myself.

2. Gilbert
   Comment by Gilbert | 06/11/08 at 7:00 pm

   Good howto, BUT! don’t forget to omit the conf dir as different servers means different IP’s… and chroot dir. can also be excluded. –exclude */conf/* –exclude chroot/*.

   this applies when a plesk migration manager has been used prior to rsync.

   HTH

3. Corey
   Comment by Corey | 08/06/08 at 10:10 pm

   Um, rather than backing up live MySQL files that are susceptible to corruption, I do mysqldump against every database on a given hosting server. Something like so:

   find /var/lib/mysql -type d | sed -e ’s/\/var\/lib\/mysql\///’ | \
   grep -v “/var/lib/mysql” | \
   xargs -i \
   mysqldump -u admin –password=`cat /etc/psa/.psa.shadow` -r /var/dump/mysql/{}.sql {}

   Basically I created /var/dump/mysql and each night, prior to my rdiff-backup job, I overwrite the SQL dump files. rdiff-backup then keeps up to 7 versions of the file until they are pruned due to age.

   Not only does that protect you from corrupt database files, it also saves you if site A needs to do a restore but site B is ok, etc.

4. erik
   Comment by erik | 08/07/08 at 9:03 am

   Hi Corey, actually I do the same as you these days. I make full database dumps and then run rdiff-backup over the entire server. It’s great to be able to restore a database or a file from x days ago.

5. Alexander Driantsov
   Comment by Alexander Driantsov | 09/02/08 at 2:09 pm

   I have hardcoded the script like this, for our hosting(it’s working perfectly, sending very good with load average and time taken for each user’s webroot backup):

   —————- CUT —————–
   #!/bin/sh
   # Incremental hosting backup script using rdiff
   # Author: Driantsov Alexander
   ####

   BACKUP_ADMIN_EMAIL=”my-email@my-host.com”
   BACKUP_MAIL_SUBJECT=”`hostname`: Backup Synchronization Result”
   BACKUP_LOG=”/var/log/backup-log”

   BACKUP_DELAY=”10″ # Delay between clients backup(seconds)
   BACKUP_PRIO=”20″ # Priority for the MySQL dump and rdiff-backup Min: 20 Max: -20
   BACKUP_DIFF_DIR=”/var/backup/hosting” # Diffs of dumps will be stored there
   BACKUP_TTL=”6M” # Time To Live

   SYNC_PRIO=”20″
   SYNC_SRV=”254.254.254.254″ # Remote server for backup storage
   SYNC_USER=”backup-user” # User at remote storage
   SYNC_SPEED=”200″ # Limit Synchronization Bandwidth to this number of KB/s
   SYNC_DIR=”/backup/hosting” #Directory on Remote server to synchronize backups in

   EXCLUDE_ARRAY=”**templates_c
   **html_cache”
   TMP_DIR=”/tmp”

   echo -en “$EXCLUDE_ARRAY\n”|while read exlude_dir; do exclude_str=`echo -en “${exclude_str}–exclude ${exlude_dir} “`; echo $exclude_str > $TMP_DIR/exclude_str.tmp ; done
   EXCLUDE_STR=`cat $TMP_DIR/exclude_str.tmp`
   rm $TMP_DIR/exclude_str.tmp

   backup_start=”$(date +%s)”
   load_average=`uptime|awk ‘{print $10″ “$11″ “$12}’`
   echo “Load overage at start: $load_average” > $BACKUP_LOG
   echo “BackingUP WebRoots:” >> $BACKUP_LOG
   mysql -uadmin -p`cat /etc/psa/.psa.shadow` -Bse “select login,home from sys_users;” psa| while read user path; do
   before_backup=”$(date +%s)”
   BACKUP_RESULT=`nice -n $BACKUP_PRIO rdiff-backup $EXCLUDE_STR $path/httpdocs $BACKUP_DIFF_DIR/$HOSTNAME/$user 2>&1`
   after_backup=”$(date +%s)”
   elapsed_seconds=”$(expr $after_backup - $before_backup)”
   load_average=`uptime|awk ‘{print $10}’|sed -e ’s/,//’`
   if [ -n “$BACKUP_RESULT” ]; then
   echo -en “$load_average - $elapsed_seconds secs - $user - ERRS: $BACKUP_RESULT” >> $BACKUP_LOG;
   else
   echo “$load_average - $elapsed_seconds secs - $user” >> $BACKUP_LOG;
   fi
   load_average=`uptime|awk ‘{print $10″ “$11″ “$12}’`
   sleep $BACKUP_DELAY;
   done

   backup_end=”$(date +%s)”
   backup_elapsed_seconds=”$(expr $backup_end - $backup_start)”

   echo “\nLoad overage after WebRoots backup: $load_average\n” >> $BACKUP_LOG
   #echo “Time taken for backup: $elapsed_seconds seconds” >> $BACKUP_LOG

   #Synchronization
   sync_start=”$(date +%s)”
   data_result=`nice -n $BACKUP_PRIO rsync -avz –bwlimit $SYNC_SPEED $BACKUP_DIFF_DIR/$HOSTNAME $SYNC_USER@$SYNC_SRV:$SYNC_DIR/$HOSTNAME|tail -n 2 2>&1`
   sync_end=”$(date +%s)”
   sync_elapsed_seconds=”$(expr $sync_end - $sync_start)”
   free_space_info=`nice -n $BACKUP_PRIO ssh $SYNC_USER@$SYNC_SRV df -h –sync -t ext3 2>&1`

   echo -en “Time taken for backup: $backup_elapsed_seconds seconds\nTime taken for synchronization: $sync_elapsed_seconds seconds\nWebRoot backup synchronization:\n $data_result \n\nInformation on free space on remote backup server: \n $free_space_info \n\n Backup Log: \n\n`cat $BACKUP_LOG`”| mail -s “$BACKUP_MAIL_SUBJECT” $BACKUP_ADMIN_EMAIL
   —————- CUT —————–

   Additionally we have such script for MySQL backups, these scripts works perfectly with load average <= 1. Good luck in codind and fighting with overheads caused by backups, I spent a days to findout the best way of MySQL backup, trying mysqldump, mysqlhotcopy and other methods

Leave a Reply »»

Name

Email

Website

• Comment Preview