Re: [Duplicity-talk] Incremental backups added onto a segfaulted full backup

[edgar . soldin]

On 24.04.2015 22:15, E.B. wrote:
>>> A full backup process had a segfault many days back. The process left
>>> behind a lock file but after the lockfile was deleted, subsequent 
>>> incrementals
>>> have succeeded normally. Is data in this backup chain reliable? What is best
>>> course of action? Delete and do a full asap?
>>
>> do a verify on that chain. if that fails, consider the chain corrupt and do 
>> a new full.
> 
> Sorry, a couple newbie q's, I hope you can indulge me
> 
> - does a verify pull data from the remote? if so, how much? the whole shebang?

verify is a local restore of the whole chain. so, yes, one whole chain.

> 
> - what's the process to delete a chain from local and remote (when I have 
> other
> chains both before and after it that should be saved)?

you do not have local chains. all backups are stored remote, while locally 
metadata of them is kept (archive-dir), which can be recreated if out of sync 
with remote.

you cannot selectively delete backups. command purge works for a timedate or a 
number of fulls.. see manpage for details.

having said that, the backup file structure is pretty forward. files are named 
like what they are and have a timestamp in their name. have a look and you will 
probably identify the whole broken chain and it's files. delete them and you've 
'selectively' deleted it.

..ede/duply.net