Re: [Duplicity-talk] Restart is impossible

[edgar . soldin]

On 14.05.2012 21:23, Tim Eggleston wrote:
> Hi list,
> 
> On attempting to back up a large (250GB) fileshare area, duplicity gets 
> interrupted due to a problematic internet connection. It is unable to resume 
> the backup from where it left off, giving the message:
> 
> RESTART: The first volume failed to upload before termination.
>          Restart is impossible...starting backup from beginning.
> 
> And then starts back again at volume 1.
> 
> However, looking on the backup target server I can see that many of the 
> volumes have in fact been uploaded successfully.
> 
> On smaller areas, full + incremental backups work fine, and restores work 
> fine. It's just the restart functionality which is problematic.
> 
> This is a Debian box backing up to a Ubuntu box, running Duplicity 0.6.18. My 
> invocation is as follows:
> 
> duplicity full --encrypt-key "${KEYID}" --sign-key "${KEYID}" --num-retries 3 
> --name fileshare_files --exclude /mnt/files/lost+found /mnt/files 
> scp://address@hidden//data/backups/${CLIENTNAME}/duplicity/fileshare_files"
> 
> Am I doing anything obviously wrong? Could this be a bug?
> 

i would suggest to abstain from having duplicity to upload for backups that big 
and rather do the backup to a file:// target and upload/sync this manually.

resuming is still not 100 percent and seems to fail sometimes.

what i don't understand though. you limit the retry count to 3 although having 
interrupted uploads. raise that number so duplicity has the possibility to 
retry more often in the hope that the upload finally goes through.

a side note: in 0.6.18 ssh backend retrying does not work (i am working on it). 
so either:
1. use 0.6.17 (has a memory leak so your system might start to swap. you could 
add the patch manually though)
2. replace 0.6.18 ssh backend with 0.6.17's one (needs backends/ssh_backend.py 
and expect.py)
3. wait for the next release


..ede/duply.net