How do you backup?

[MrCrispy]

For those of you with lots of data, what backup strategy do you use? I have some doubts about the best way to do this.

 

1. Lets say I build a 2nd server for backup. But if the backup is realtime then you still have the sync problem, i.e. deletions are also propagated. Plus the added expense and running costs of a 2nd server.

 

2. Using external disks that you attach and backup manually. The problem here is managing what goes on what disks, keeping track of what's full etc. This is how I have my current media stored and it becomes a pain

 

3. Duplication of data (e.g. Windows DrivePool). Seems to me like a good compromise, you have 2 copies of data, its not a true backup but even #1 has its limitations.

 

I realize what I'm discussing here is not true backup (ideally I'd have offsite backup) but that's simply not possible due to upload bandwidth limitations. And I don't have anyplace to keep an offsite server anyway.

 

Thoughts?

 

[bjoernwk]

Good morning,

 

we are using crashplan and a second unraid machine. This scenario is free if you do not use the option, to backup to crashplan-cloud servers.

 

Our most important data is additonally sent to crashplan. Of course, this is impossible with all the multimedia stuff, but our photos and documents and music is sent off site. This costs about 15 $ per month (and includes 10 machines to backup with no limit to the cloud). Crashplan supports versioning, so if a file is corrupt, you can replace it with an different version from the day before.

Crashplan integrates perfectly in unRAID with the APP. This is our backup-solution and I sleep very well at night. : - )

 

Greetings from Germany!

 

Björn

[Fireball3]

Using rsync to mirror my main server to a backup machine.

 

I would use crashplan but I don't want to set up dockers just for this one and only function.

If there was a crashplan plugin I would use it.

[bjoernwk]

Good point, but I am using dockers anyway, so no problem for me. Just for CP I wouldn´t use it neither.

 

Rsync is a mighty way to do this. But I would miss the versioning feature. Or is there a way to do this with Rsync as well? I am not familiar with it.

[Fireball3]

Yes, rsync can do versioning also.

In fact, rsync is a pretty mighty tool. The only problem is it's missing GUI.

Hacking the console each time is pretty geeky and will hold off many users.

[Naldinho]

I'm still working on a backup plan but this is the general outline so far.

 

1) Media -- just relying on unRaid having the one disk failure tolerance.

 

2) Essential Stuff -- Crashplan backs up office stuff to Home unRaid and home stuff to the unRaid machine at the office -- thinking of adding a third version. If I add a third version it would be to a computer on its own network at home with absolutely nothing else on it. The home and office networks are connected by a site-to-site VPN and I don't know if all this talk about ransom-ware propagating across networks is just hype/paranoia or an actual threat.

 

3) Windows machines (physical and VMs) -- all the user files including desktop have had their destinations changed to folders on the local unRaid. Those folders get backed up as part of the essential stuff backups so restoring is just doing a clean reinstall of all the software. Next time I need to do a clean install I'll start making a snapshot / image of the clean brand new system install but I'm not sure how much time that would save over just a clean install.

 

 

Stuff I need to figure out still

 

1) Websites with databases -- currently rely on hosting service to back up with shared hosting but will be moving to unmanaged VPS in the next couple of weeks so need a way to back those up to unRaid -- preferably automated and with versioning.

 

2) Linux VMs -- two of them. One is my development machine where i work on stuff before deploying on live sites and the second is a internal webserver where I run a web-based CRM/ERP I'm developing that basically will eventually have all the essential business data.

 

3) PostgreSQL database that is currently the only thing that is not out of the box software on two windows machines. The two machines independently maintain their own copies and all the data required to rebuild the database in a few hours is backed-up but I'd prefer a solution where there is one database only (sync issues) used by two computers so thinking the Postgre docker.

 

4) Android devices -- phones and tablets -- currently nothing but I should at least consider something for photos on my phone.

[Fireball3]

With regard to ransomware maybe this gives you some insights.

https://lime-technology.com/forum/index.php?topic=50949.msg489292#msg489292

[Naldinho]

Thanks. Interesting and scary read. My backups are on different subnets in different physical locations but if a Windows machine can browse to the shares it can get to them so  I think I am going to add a third backup.

 

I was thinking of doing something with the third backup only connects when it is backing up. I don't think that is actually necessary since it is part of a different network but it is easy to setup so why not? 

[MrCrispy]

My truly essential data is backed up to Google Cloud, OneDrive and 2 different USBs, but I'm reaching size limits. e.g. for photos right now I have a GDrive promo that's set to expire so I'll run out of space.

 

Everything else right now is managed manually by juggling externals and drives in my windows pc. Using disk catalog programs and FreeFileSync it works reasonably well and I actually have most of my data duplicated, but its a lot of work.

 

[matt_webb]

Crashplan for me.  Got 2 usb drives and swap them between work and home every week or so.  I also use Crashplan's online backup.  Works well.

 

Sent from my SM-N920I using Tapatalk

 

 

[AndrewT]

1. Lets say I build a 2nd server for backup. But if the backup is realtime then you still have the sync problem, i.e. deletions are also propagated. Plus the added expense and running costs of a 2nd server.

 

As others have noted, fancy backup solutions like CrashPlan allow you to hold onto files for a given time even if your path your backing up now has the file removed. Rsync is nice and is what I use mostly but I don't know of an easy solution to the feature of holding onto a file that was deleted within the past year (as an example).

 

Here's what I use for rsync to preserve time stamps and permissions:

rsync -lrtuvD <current-path-to-backup> <backup-path>

 

I do that manually once or twice a year with files especially important to me.

[bjoernwk]

Man, I think I should invest more time in getting familiar with Rsync. Thanks a lot for the ideas.

 

Greetings from Germany!