Services down during backup?

IP address limiting won’t help much if the device is compromised.

The down time seems to be from transferring huge files across the Internet by using the optional external backup location which I believe is configured in the dashboard and keeps the services from being restarted. If you don’t configure it, then the local operation is just the speed of the server, likely just minutes.

This is just silly. If you are using as backup, of course you need to ensure that “device” is a secure medium as well e.g. set up another VPS that have all your security set up, or use your home server but make sure the same security is applied.

If that’s what is happening then it probably explain why most of us did not see a long down time. We have most likely backuped to a location, with speedy tranfer. e.g. rsync to another vps on same network, or an attached volume.

Why attached volume / elastic volume

  • it’s cheap.
  • it’s fast
  • if your server crashes, your volume is not affected.
  • you can remount the volume to another server if your server crashes

Then again my www-data is already on an elastic volume… headache, i think i have too many back up solutions in place. manual rsync, elastic volume, s3.

Have you found a solution yet? You mentioned:

This should be your solution, although I would approach it differently than suggested earlier in this thread.

I started a similar thread several weeks ago: Sever software shut down in the middle of the night (must not have seen this one).

I’m not convinced the network copy is totally responsible for the length of time the backups are taking. My first backup took nine hours for 150gb. This was just going to local disk, no network copy was involved. As a wild guess, I think maybe the encryption and or compression is responsible.

As a comparison, the same data on my old server took about two hours to send compressed tar files to another in-house server. This was done from an LV snapshot of the volume so there was no down-time.

Has anyone tried creating a snapshot on an MaiB box?

Recently scanned through the web and found find some post about duplicity being slow, when backing up large folders, especially if you have 150gb emails, which likely contains many many small files.

Each incremental update will need duplicity to scan through all the files, the lesser memory you have the slower it will be it seems.

If things are not going well with duplicity because of your large volumes, then probably a manual solution is needed, then disable duplicity by disabling miab backup.