Re: [Duplicity-talk] Local cache directory with hashes at ~/.cache despite archive-dir and name set

[Cláudio Gil]

Hi,

At least since 0.6.26, the combination of --archive-dir and --name work as expected. The default for --archive-dir is ~/.cache/duplicity and the default for --name is a hash of the destination URL. It might be easier to remember "archive dir" as "cache dir".

So if you specify both you get no surprises, when changing the destination (backend) URL. For example:

duplicity full --archive-dir ./cache --name test ./data s3://...

Will always (at least since 0.6.26+) keep a cache in ./data/test and never use ~/.cache. This will remain true regardless of the last URL. Obviously, if you change the destination duplicity will synchronize the local cache as that is how it works.

Regards,
Cláudio

Em 03/01/2017 08:09, "imagenesis--- via Duplicity-talk" <address@hidden> escreveu:

This is on 0.6.22. If this behavior has not changed, I believe this behavior is inconsistent with any logic, especially hashing against the destination and should be changed. In addition, I believe if this implicit cache directory is necessary in addition to archive-dir, it should use names when they are provided instead of hashes. 

On Tue, Jan 3, 2017 at 3:04 AM, address@hidden <address@hidden> wrote:

I have a local directory that appears to be a copy of the directory that I specify as my archive-dir except that it appears that the hashes correlate to --name and the backup destination folder. 

As specified, the back destination folder seems to be involved with generating the hash corresponding to a named backup. 

I have 3 questions, why is the backup destination folder involved?

1. I changed the path of the backup destination folder and the hash folders corresponding to the named backups had to be recreated. Is there a way to avoid having the backup destination folder involved with generating the hashes so this cache directory would not have to be repopulated when the destination path changes?

2. Why is this cache directory in use when it is specified that it is only in use when an archive-dir is not specified?

3. It would appear that this directory may have something to do with resumable backups incase of network failure. If so, the documentation is either incorrect with regards to --archive-dir or confusing. Please specify the nature and functionality of the ~/.cache directory when archive-dir is specified. 

4. Is it possible to implement repopulation of this cache directory from the -archive-dir instead of the destination metadata? 

_______________________________________________
Duplicity-talk mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/duplicity-talk