• This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn more.

What is your backup plan?

Razasharp

Well-known member
#3
Off site back-up plan with nimbus hosting.
Is that automatically done by your hosting company Robby? Is it another server in the same datacenter?

My current set up is two drives in raid mirror, and remote back-ups to a second server (just DBs) and another to Amazon AWS (DBs daily and triple sync on the files)
 

Robby

Well-known member
#4
Is that automatically done by your hosting company Robby? Is it another server in the same datacenter?

My current set up is two drives in raid mirror, and remote back-ups to a second server (just DBs) and another to Amazon AWS (DBs daily and triple sync on the files)
Yep, thats right. My sites are backuped in anothet datacenter so you have easy of mind.
 

AndyB

Well-known member
#5
My forum details:

1.3 million posts
350K attachments
Dedicated Server on CentOS with Webmin

My backup details:
  1. I have two hard drives in the server.
  2. The second hard drive is strictly for backup of the database and attachments.
  3. I have about a dozen cron jobs created with Webmin to backup and rsync the database and attachments to the backup drive. Many of the cron jobs deal with making a new backup for each day of the week.
In addition I also will rsync my database and attachments from my server to my home iMac external drive. This is done every month or so for addition peace of mind.
 

Razasharp

Well-known member
#6
Yep, thats right. My sites are backuped in anothet datacenter so you have easy of mind.
Hi Robby, I don't mean to scare you, but I've heard many stories of datacenter backups failing - I would strongly advise an independent backup to a third party.

My forum details:

1.3 million posts
350K attachments
Dedicated Server on CentOS with Webmin

My backup details:
  1. I have two hard drives in the server.
  2. The second hard drive is strictly for backup of the database and attachments.
  3. I have about a dozen cron jobs created with Webmin to backup and rsync the database and attachments to the backup drive. Many of the cron jobs deal with making a new backup for each day of the week.
In addition I also will rsync my database and attachments from my server to my home iMac external drive. This is done every month or so for addition peace of mind.
Sounds like you a good plan Andy!

Have you thought about setting up the drives in Raid so that you automatically get a copy of your data? (Making it easy to rectify if one fails). That's my current set up, but I also backup to a remote server as mentioned above.

I like using something like AWS (offsite) in addition to normal backups (on the same server), because once set up you don't really have to worry about anything. I think you can get a free year too.
 

Razasharp

Well-known member
#8
I prefer the simple solution of having a dedicated backup drive.
The main reason I go with a raid mirror is to minimise downtime in case of a hard-drive failure, and with 'some' benefit of it being a back (but my proper backups are separate). Hopefully making it easier than reinstalling OS, and setting everything up again :)
 

AndyB

Well-known member
#9
The main reason I go with a raid mirror is to minimise downtime in case of a hard-drive failure, and with 'some' benefit of it being a back (but my proper backups are separate). Hopefully making it easier than reinstalling OS, and setting everything up again :)
Hard drives are very reliable. So I'm not terribly concerned about having to replace the main drive. I plan on upgrading my dedicated server every 5 years or so anyway.
 

Tracy Perry

Well-known member
#10
Hard drives are very reliable. So I'm not terribly concerned about having to replace the main drive. I plan on upgrading my dedicated server every 5 years or so anyway.
Yeah? Try telling that to the two drives that failed in the ImageGrid PACS server at the local Dr. office I help out at. ;)
Even worse, even with an Acronis image of the server (that was done before the drive failure) we can't move it to new hardware since it is configured as an "appliance" and bonded to the hardware. :cautious:
Luckily it was just replace the drives, recreate the drive array since we replaced the controller card also just to be on safe side (was RAID 5 and had 4 drives) and then burn the Acronis image back to it... but it still took several hours.
 

Razasharp

Well-known member
#11
Yeah? Try telling that to the two drives that failed in the ImageGrid PACS server at the local Dr. office I help out at. ;)
Even worse, even with an Acronis image of the server (that was done before the drive failure) we can't move it to new hardware since it is configured as an "appliance" and bonded to the hardware. :cautious:
Luckily it was just replace the drives, recreate the drive array since we replaced the controller card also just to be on safe side (was RAID 5 and had 4 drives) and then burn the Acronis image back to it... but it still took several hours.
Sounds like a nightmare!

Although hard drives are usually pretty reliable, they do fail - and the chances of failure increase in a server where they are constantly on the go. This is why I prefer the fail-safe of raid, in the event of one failing another can be added and the array rebuilt. Chances of two failing at the same time should be slim - unless you are as unlucky as your local Dr's office :p
 

Tracy Perry

Well-known member
#12
Although hard drives are usually pretty reliable, they do fail - and the chances of failure increase in a server where they are constantly on the go. This is why I prefer the fail-safe of raid, in the event of one failing another can be added and the array rebuilt. Chances of two failing at the same time should be slim - unless you are as unlucky as your local Dr's office :p
I don't think we've had a RAID failure on any of the servers with SAS drives in over 10 years. I know that we did have 1 (out of 9) Seagate drive fail on a Gateway server - but it wasn't that big of a deal since it was call Gateway and they had the new drive to us the next day.
SATA on the other hand I don't trust any further than I can throw them. Both the drives that failed were WD 250GB Enterprise class (Dell labeled).
 

SchmitzIT

Well-known member
#15
My backup details:
  1. I have two hard drives in the server.
  2. The second hard drive is strictly for backup of the database and attachments.
  3. I have about a dozen cron jobs created with Webmin to backup and rsync the database and attachments to the backup drive. Many of the cron jobs deal with making a new backup for each day of the week.
In addition I also will rsync my database and attachments from my server to my home iMac external drive. This is done every month or so for addition peace of mind.
Would you be willing to share the script you use? I'm not overly familiar with Linux, but have been looking for something like this for a while. Anything to get me started would be nice :) Also, how do you synch it to your home drive? Is that something you do manually, or is it all automated?


One of my favorite ways to keep a backup of a server (OS) is to use a RAID 1 set with 3 HDS. Install the OS and all programs on it, then pull ont of the HDs out and stick the spare one in. The one you pulled out is your backup; the other two will now be the active mirrored set. In case you need to rebuild, simply remove the existing drives and stuff in the backup drive. Insta-OS-restore.
 

Mouth

Well-known member
#16
I use backup-manager and automysqlbackup to provide a local daily archives, then upload the archives to S3.
For additional safety and security, I have have a backupsy server utilising rsnapshot
 

woei

Well-known member
#17
I make daily backups with cPanel and transfer them automatically to my NAS at home so they are offsite. Before I used a separate drive to write te backups to and once a month I would download them manually. Now I have automated it to daily :)