Only sync relevant remote metadata to prevent big cache

Bug #1702715 reported by Markus
16
This bug affects 3 people
Affects Status Importance Assigned to Milestone
Duplicity
New
Undecided
Unassigned

Bug Description

Duplicity version: 0.7.13.1
Python version: 2.7.12
OS Distro and version: Ubuntu 16.04.2
Type of target filesystem: Multiple

Hello,

this is a follow up of http://lists.nongnu.org/archive/html/duplicity-talk/2017-07/msg00000.html (slightly modified).

I've been using duplicity to backup since 2014 with a full backup every 3 months, so I have some secondary backup chains. Additionally backups go to multiple separate places (each having multiple backup chains). My local cache folder is now over 70gb big and I had the idea to delete the cache for the secondary chains to free up space - since they are only needed when restoring files, but not for new backups (to the primary chains).

So I deleted the cache files (for example duplicity-full-signatures.20141106T121637Z.sigtar.gpg) and reclaimed a lot of space.

But after the next backup run, I saw this in the logs:

Synchronizing remote metadata to local cache...
Copying duplicity-full-signatures.20141106T121637Z.sigtar.gpg to local cache.
Copying duplicity-full-signatures.20150505T234205Z.sigtar.gpg to local cache.
Copying duplicity-full-signatures.20150904T234123Z.sigtar.gpg to local cache.
Copying duplicity-full-signatures.20151205T004128Z.sigtar.gpg to local cache.
...

My current workaround is: Before I create a new full backup, I move the remote backup files to a new location. So for duplicity it looks like the backup location is fresh and it even deletes the local cache automatically, starting with a full backup (which I wanted anyways).

If I later want to restore from that backup I will have to move the files back into place on that remote.

It would be great to have a command line option to prevent syncing all metadata but the current one. Maybe it could even determine which metadata is needed for the current operation and only sync this one, so for example if I wanted to restore from a secondary backup chain it would only sync the secondary backup chain's metadata. So if metadata was deleted manually but not needed for the current operation, it would stay deleted.

Tags: cache metadata
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.