New Veeam user pulling hair out. Backup copies for offsite..

[adessmith]

Hey everyone,
I'm a new Veeam user and trying to figure out my offsite backup schedule. Sorry if this turns out to be a long post.
I'm not sure I can explain everything I have tried, and the issues I have encountered, as I have tried so many things and had so many issues I have confused myself.
I will try to explain what I have and what I would like to do. Hopefully someone here can give me a recommendation as to how to set this up, and explain to me where I have gone wrong.

I have Veeam B&R 8 essentials standard installed on a machine with a dedicated raid 1 local repository on a pair of 1TB drives.
I have 2 ESXI 5.5 hosts running a combined total of 10 VMs. I have all my regular backups to the local repository working exactly as I want:
Job 1 - Does a regular incremental backup for 9 VMs at 3PM every day, with a full backup on Saturdays.
Job 2 - Does a regular incremental backup for 1 VM at 5:05PM every day, with a full backup on Sundays.

This all works great, but now I am trying to work out my offsite strategy.
Due to budget constraints, backup sizes, available bandwidth, and connection reliability, cloud backup or off site backup / replication over our internet connection is not an option at this time.
I have a pair of portable rugged 1TB USB 3.0 drives I need to swap out each day, taking one off site when I leave.
The business closes at 5PM each weekday. I leave at roughly 5:30, which gives plenty of time for an incremental backup.
When I have made this work it seems like the backup window (between 5:05 and 5:30) has been plenty for all this to happen. I just have had issues with setting up the schedule to automate it all.
The one backup that runs at 5:05 is a SQL server which contains business transaction data. I run a SQL maintenance job over the weekend which causes the incremental backup to be much larger, so I have scheduled it so that my full VM backup occurs a few hours after my SQL maintenance plan.

I thought I had it all worked out before I left last night. The two regular backup jobs ran at 3PM and 5:05 respectively.The backup copy jobs saw changes to those backups, and copied them to my offsite drive immediately after each backup job... just as I expected. Before I left I swapped the drives, and left for the night carrying the one I just unplugged off site and leaving the other plugged in to capture the next day's backup copy.
I get home and started getting emails all night that my offsite backup (the one for the 9 VMs) had failed. I believe it was one email every hour. I had it set to copy every "1 hour", but doesn't that just mean that it would check for changes to the backup file every hour, and do nothing if there was no change? No other backup jobs ran after I left last night... so I don't understand why the backup copy job was trying to do anything. I had the other job setup the same way, but didn't receive emails about it.

Here is something I don't get:
Other than the time and VMs specified, both backup jobs and both backup copy jobs, are configured identically... The current status of both copy jobs is "idle", but when I checked this morning one copy job was sitting at "waiting for the new copy interval", and the other was sitting at "waiting for new restore points". I would expect them to be both "waiting for new restore points".
I just disabled both jobs, and re-enabled them, and now they both display "waiting for new restore points".

Before yesterday I only had 1 USB drive, and I wasn't sure what to set the "copy every" setting to on the copy job.
I thought I needed this set to copy every "1 day", but it seemed to cause problems with my full backup. The full backups would run over the weekend, and when I plugged the drive in Monday morning the two jobs would start to make copies.
Then at 3 and 5:05 it wouldn't create copies, presumably because it was set to only make 1 copy each day, which already happened.

Now that I have 2 drives, this seems to complicate things further. I have enabled the rotated drives option on the repository, but I am confused on how this works and plays into my backup strategy.

Another thing I don't understand is the "cycle" for copy jobs.
If I set the copy job to copy every "1 day", does that reset every 24 hours from the time the cycle was started, or does it run from 12AM to 11:59PM?
For instance:
Lets say the copy job cycle was started at some point at 10AM and set to copy every "1 day".
My backup job executes Sunday at 5:05PM and does a full backup.
Monday morning at 8AM I plug in a drive and it copies the backup.
Monday evening at 5:05PM an incremental backup occurs.
When will the next backup copy happen? Does it happen immediately because there has not been a backup taken since 10AM, or will it wait until midnight (the following day)?

Any help here would be appreciated.

[foggy]

I had it set to copy every "1 hour", but doesn't that just mean that it would check for changes to the backup file every hour, and do nothing if there was no change?

No, backup copy job checks for available restore points automatically. Synchronization interval is the time range where it looks for restore points and there's not much sense in setting it more frequent than the actual backup.

Other than the time and VMs specified, both backup jobs and both backup copy jobs, are configured identically... The current status of both copy jobs is "idle", but when I checked this morning one copy job was sitting at "waiting for the new copy interval", and the other was sitting at "waiting for new restore points". I would expect them to be both "waiting for new restore points".

Not sure why the second job behaved differently, but looks like it hadn't restart synchronization interval, hence different idle message.

Before yesterday I only had 1 USB drive, and I wasn't sure what to set the "copy every" setting to on the copy job.
I thought I needed this set to copy every "1 day", but it seemed to cause problems with my full backup. The full backups would run over the weekend, and when I plugged the drive in Monday morning the two jobs would start to make copies.
Then at 3 and 5:05 it wouldn't create copies, presumably because it was set to only make 1 copy each day, which already happened.

Now that I have 2 drives, this seems to complicate things further. I have enabled the rotated drives option on the repository, but I am confused on how this works and plays into my backup strategy.

Backup copy job copies one latest VM state per sync interval, so there should not be any issues with weekend fulls (corresponding restore point will be created on disk plugged in on Friday).

If I set the copy job to copy every "1 day", does that reset every 24 hours from the time the cycle was started, or does it run from 12AM to 11:59PM?

From the time you specify as the copy interval start. I recommend reviewing the corresponding user guide section. If you still have any questions regarding how sync interval works, feel free to ask.

[adessmith]

Not sure why the second job behaved differently, but looks like it hadn't restart synchronization interval, hence different idle message.

Now that I think about it, the only thing that makes sense is that I changed the sync interval on both jobs prior to the backup... maybe I didn't allow one of the jobs to restart with the new settings...

Backup copy job copies one latest VM state per sync interval, so there should not be any issues with weekend fulls (corresponding restore point will be created on disk plugged in on Friday).

I think this makes sense to me now... The main thing I am concerned about at this point is how the rotated drives play into the equation. I'm not sure I understand what happens when you swap out drives...
Let's say my full backup happens on sunday night on drive A, the next time I remove that drive and replace it with drive B, won't that cause it to copy a ton of data over again? does it wait until the next sync interval before it sees that drive B doesn't have the previous backup and copy the missing data, or will it do that ahead of time? I wan't my weekly backups to run between close of business and 5:30 during the weekends, but I think this is going to cause the first drive swap after the full backup to take considerably longer...

[foggy]

Drive swap itself does not trigger data transfer, the job always waits until the new interval starts and looks for the new restore point. In case new empty drive is inserted, it will copy the data required to build a full backup and will create increments after subsequent drive swaps.