[Borgbackup] VM backup issues

[Ken Bass]

On 11/16/2017 08:30 PM, Thomas Waldmann wrote:
>>> You are using mtime of a regular (disk image) file or the LVM device
>>> file?
>> Both depending on what the underlying file is... os.stat following
>> symlinks will indicate that the mtime of either. That appears to be what
>> I want.
> Hmm, is there a strict relationship between the mtime/ctime of the
> device file and the device contents?
Sorry I am not sure of your question here. I know in my use case if a VM 
has never been powered on since the last backup, the mtime/ctime of the 
underlying LVM will not have been touched, so there is no need to chunk 
a 30G or 100G file just for the dedup to ignore it. In my case it is a 
waste of time / power.

>
>> Correct - it was just a typo on my part, I should have used
>> 'glob.glob('*')' but I thought the result was inconsistent - making an
>> empty backup. In fact is seems downright dangerous to proceed with an
>> empty backup. If someone were to do a prune afterwards they could
>> potentially wipe out their last good backup depending on how many they
>> are keeping.
> That has nothing to do with empty or not empty.
> The question is if the archive contains what you specified.
>
> Missing one important file or folder can be even worse than being empty
> because it is not that obvious.
>
> And borg indicates that there was something going wrong via rc 1 and a
> warning.

I understand what you are saying, but I  think specifying a set of files 
that are ALL errors should return an rc of 2 and not create an empty 
backup. Just my opinion about what is less error prone and more user 
friendly.

My automated script did see that was a non zero error code because an 
exception was thrown. I just didn't expect an empty data set to be 
present from that run.

>> There is no description of what any of the returns code are.
> http://borgbackup.readthedocs.io/en/stable/usage/general.html#return-codes
>
> There is a table that explains it.
>
> I also see the table in the man page.
>
I am running the rpm's from EPEL on Centos. Those are the man pages 
missing that info. Thanks for the pointer to the documentation page--I 
missed that. (I was probably looking under the create command and missed 
the general/common usage area)