[Borgbackup] migrating from attic

[Thomas Waldmann]

>> Sure, as it copies your repo, space requirement doubles.
> http://borgbackup.readthedocs.org/en/stable/usage.html#borg-upgrade
> "Unless --inplace is specified, the upgrade process first creates a
> backup copy of the repository, in REPOSITORY.upgrade-DATETIME, using
> hardlinks."
> I thought that hardlinks part meant it wouldn't use twice the space...

Well, this seems a case of too much (misleading) details in the docs.

Yes, it uses hardlinks at first (to quickly make a pseudo-copy of the 
repo). This does not need much space.

The reason why it does this is via hardlinks is because it is quick.

After that, it will iterate over all files, show a progress indication, 
read, convert and rewrite them (thus breaking the hardlink). And while 
it does that, space requirements doubles.

>> Did you run it as the same user as the attic backup and as the
>> migration?
> I did migration locally, but backups was made as the same user remotely.
>
>>
>> IIRC the migration converts the attic chunks and files cache, so
>> quick skipping of unmodified files and deduplication should work.
> Is there a way to fix it? I can't add space to existing backup storage.
> But I can copy backup repo to some other place, make space consuming
> operations and copy it back.

I'm still not sure what went wrong there. The cache is located via the 
repo-id, so it should match no matter where that repo dir is located.

Did you run the repo conversion on the machine that is normally running 
the borg (client) to backup the data that is in the repo?

-- 

GPG ID: FAF7B393
GPG FP: 6D5B EF9A DD20 7580 5747 B70F 9F88 FB52 FAF7 B393