duplicity-talk
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Duplicity-talk] is a duplicity folder rsync-friendly?


From: edgar . soldin
Subject: Re: [Duplicity-talk] is a duplicity folder rsync-friendly?
Date: Mon, 12 Aug 2013 16:38:10 +0200
User-agent: Mozilla/5.0 (Windows NT 5.1; rv:17.0) Gecko/20130801 Thunderbird/17.0.8

On 12.08.2013 16:28, Martin Erik Werner wrote:
> On Mon, Aug 12, 2013 at 04:05:09PM +0200, address@hidden wrote:
>> On 12.08.2013 15:53, Martin Erik Werner wrote:
>>> Hi,
>>>
>>> I'm having problems using duplicity with a cifs mount, when I try to
>>> make a backup to cifs it crashes with "CIFS VFS: No writable handles for
>>> inode", I'm seeing similar issues when I try to "git push --mirror" onto
>>> the mount as well.
>>>
>>> However, straight up copying seems to work ok, hence I'm wondering if it
>>> is feasible to create the duplicity archive locally and then rsync it
>>> over to the cifs mount. Provided I don't make any full backups, is the
>>> duplicity on-disk structure in itself incremental? Such that this method
>>> would work and be as space-efficient as running duplicity directly to
>>> the dir?
>>>
>>>   $ cp -r /mnt/cifs/duplicity ~/duplicity
>>>   $ duplicity ~/important/ file://duplicity
>>>   $ rsync -a duplicity/ file:///mnt/cifs/duplicity
>>>
>>
>> that'll work. why shouldn't it? duplicity simply creates new files in the 
>> backup repository on each incremental. simply have a look at the folder. it 
>> has files neatly numbered with timestamps.
>>
>> but maybe you rather solve the problem, maybe it is this
>> http://lkml.indiana.edu/hypermail/linux/kernel/1210.2/03971.html ?
>> you could try switching kernel or userland binaries, depending on what way 
>> to mount cifs you use.
>>
>>
>> ..ede/duply.net
>>
> 
> Ok, good, at least that would work then.
> 
> Hmm I'm running on
> 3.2.0-4-amd64 #1 SMP Debian 3.2.41-2 x86_64 GNU/Linux
> with cifs-utils 2:5.5-1 (5.5)
> and duplicity 0.6.18
> 
> So I'm guessing that a regression in Linux 3.7-rc2 might not apply? Even
> though it's very similar to what I'm seeing...
> 

i wouldn't count on it that files transferred over it do not get corrupted. you 
should at least run some test transfers with checksumming to verify that files 
survive the transport unharmed ;)

also 
1. you could try finding some alternative kernel from the debian repo, just to 
try if it works better. even an older one would do for test purposes.
2. try alternatives like fusesmb or smbnetfs or... whatever you can find out 
there

..ede/duply.net




reply via email to

[Prev in Thread] Current Thread [Next in Thread]