Dangerous behavior when license expires (Web UI)

UX
#1

Hello!

My trial license expired at some point and my backups (understandably) started to fail. However, this was essentially a silent failure as the only place this is visible is in the Schedule tab.

Furthermore, my prune job was continuing to run successfully! So duplicacy was slowly pruning away all my old backups without adding any new ones. I don’t have these jobs set to run in parallel so I would have expected the prune job to fail if the backup job failed.

I have since purchased a license, so this is no longer an immediate issue for me, but I’m concerned about what’s going to happen when my license expires again down the road.

#2

I don’t see the issue here?

Backups expire according to rules you set and those rules generally leave the most recent backups alone until a very long time passes - i.e. what gets pruned are backups that would normally have been pruned anyway.

All non-backup jobs - pruning, copying, checking etc. - are allowed without a licence. Just make sure there’s a licence if you want to continue backups.

#3

So, the reason that I think this is dangerous is the combination of:

  1. It wasn’t immediately clear to me that my backups were failing, AND
  2. All of my backups were deleted

This left me with no off-site backups by the time I realized what was happening. My suggestions for these are:

  1. Promote job errors to the dashboard in some way. Maybe it’s just putting an :warning: on the Schedule item in the left bar? As a new user I wasn’t aware that I needed to click into Schedule to check for errors.
  2. Allow me to specify that a job failure within a schedule aborts the subsequent (non-parallel) jobs. Ex: if backup fails, do not run prune. This means that I will never be in a position where all my backups get deleted.

Thanks!

#4
  1. Agreed, this certainly needs to be more visible.

The Web Edition needs a working icon tray in the taskbar - even in service mode (which currently isn’t implemented) - to communicate activity apart from backups, such as an expired licence.

  1. Are you saying it deleted all backups including the most recent one?

From my understanding, this shouldn’t be possible - unless you were using the -exclusive flag, which shouldn’t be used in most situations.

Indeed it would be nice to have an option is abort subsequent jobs in a schedule if prior jobs failed.

#5

Hmm, you’re probably correct then about there being one backup remaining. My successful backup after entering the license was the only one I saw so I assumed it had been empty, but the storage history indicates otherwise. I’m glad to know that safety net exists!

#6

Hi, for ease of mind, also remember to enable reports to a monitoring service, to help catch problems!

When the service does not receive a “success” signal in time, a failure report can be sent to you, and also monthly reports with some stats. (Healthchecks.io at least)

GUI has built-in support for this:

…and users using CLI could script something like this:

4 Likes