Closed (works as designed)
Project:
Backup Client-Server
Version:
6.x-1.0-beta1
Component:
Code
Priority:
Normal
Category:
Bug report
Assigned:
Unassigned
Reporter:
Created:
29 Nov 2011 at 15:18 UTC
Updated:
8 Nov 2012 at 03:21 UTC
The Amazon S3 feature seems to upload website files but not database backups. I have recurring backups set up and working well for both files and the database, but only the file backups are uploaded to Amazon S3.
Comments
Comment #1
jpsalter commentedThe database dumps are located in a directory named "backup_client" within your files directory. These are included in the tar file that is uploaded to Amazon S3. So, when you un-tar the backup you will also have a copy of your database.
Comment #2
mharrisonb commentedAh ok, I understand now...I had changed the default folder for backups because the help file said the following:
"If you would like to save your files outside of your Drupal install you can use a path like this: ../backup_client. If this path is not web accessible then you'll enjoy more security with your backups."
I currently have it set to the folder ../../drupal_backup, so I'm guessing I could change that to just drupal_backup instead, and make sure that the permissions on the directory are 600 and it will be sufficiently secure. Is this correct?
Thanks for the quick response and the awesome plugin.
Comment #3
jpsalter commentedYes, placing the backup_client folder within the Drupal root (for example - the public files directory) will cause the database dumps to be included in the tar file.
I'm sorry that it isn't more clear... I've been struggling to figure out a straight forward solution but there are a lot of different possible setups to support (and in Drupal 7 we need to deal with public *and* private folders). If you have any suggestions I would love to hear them.
Comment #4
mharrisonb commentedThanks...now I'm a bit confused. On the Amazon S3 configuration page, it says:
"Each time Drupal cron is triggered - Backup Client S3 will check to see if a scheduled upload is needed. If so, the most recent backup will be uploaded to Amazon S3."
I thought that meant that when it's uploading to S3, it's not making a fresh backup but rather uploading the most recent tar.gz file (which could have been triggered by the "recurring backups" cron job at any point earlier) to S3.
I noticed that even though the drupal_backup directory I created is outside the web root, it still uploads the website backup files (the .tar.gz files) to S3. So why can't it upload the most recent sql.gz file as well?
In any case, it doesn't seem optimal to have the backup directory be within the web root because then the same backups are being uploaded to Amazon S3 again and again, which is somewhat of a waste of bandwidth.
If you can help me to better understand the issues, I'd be happy to help think of a solution. Thanks again.
Comment #5
jpsalter commentedI thought that meant that when it's uploading to S3, it's not making a fresh backup but rather uploading the most recent tar.gz file (which could have been triggered by the "recurring backups" cron job at any point earlier) to S3.
I noticed that even though the drupal_backup directory I created is outside the web root, it still uploads the website backup files (the .tar.gz files) to S3. So why can't it upload the most recent sql.gz file as well?
In any case, it doesn't seem optimal to have the backup directory be within the web root because then the same backups are being uploaded to Amazon S3 again and again, which is somewhat of a waste of bandwidth.
Comment #6
mharrisonb commentedI'm rather late getting back to this, but I moved the backup folder to be within the web root (and set the appropriate permissions on it of course), and checked the option you recommended ("Include only the most recent database backup(s) in the website backup" under "Website backup options").
I think it will do what I want now. Thanks!