Re: [Duplicity-talk] Out of space error while restoring a file

[edgar . soldin]

On 03.09.2012 11:28, address@hidden wrote:
> On 02.09.2012 09:28, Laurynas Biveinis wrote:
>> 2012/9/1 Laurynas Biveinis <address@hidden>:
>>>>>>> I'm trying to restore a ~30GB file from backups. The free space on the
>>>>>>> drive is about 80GB. Yet on restore I get the error below. What would
>>>>>>> be causing this and how much free space do I actually need to restore?
>>>>>>>
>>>>>>> $ duplicity -t2D --file-to-restore "path/to/file"
>>>>>>> --s3-european-buckets --s3-use-new-style s3+http://foo $HOME/file
>>>>>>> Local and Remote metadata are synchronized, no sync needed.
>>>>>>> Warning, found the following orphaned backup file:
>>>>>>> [duplicity-inc.20120319T102409Z.to.20120320T010946Z.manifest.part]
>>>>>>> Last full backup date: Thu Aug 16 00:00:27 2012
>>>>>>> Traceback (most recent call last):
>>>>>>>   File "/usr/bin/duplicity", line 1403, in <module>
>>>>>>>     with_tempdir(main)
>>>>>>>   File "/usr/bin/duplicity", line 1396, in with_tempdir
>>>>>>>     fn()
>>>>>>>   File "/usr/bin/duplicity", line 1330, in main
>>>>>>>     restore(col_stats)
>>>>>>>   File "/usr/bin/duplicity", line 623, in restore
>>>>>>>     restore_get_patched_rop_iter(col_stats)):
>>>>>>>   File "/usr/lib/python2.7/dist-packages/duplicity/patchdir.py", line
>>>>>>> 522, in Write_ROPaths
>>>>>>>     for ropath in rop_iter:
>>>>>>>   File "/usr/lib/python2.7/dist-packages/duplicity/patchdir.py", line
>>>>>>> 495, in integrate_patch_iters
>>>>>>>     final_ropath = patch_seq2ropath( normalize_ps( patch_seq ) )
>>>>>>>   File "/usr/lib/python2.7/dist-packages/duplicity/patchdir.py", line
>>>>>>> 475, in patch_seq2ropath
>>>>>>>     misc.copyfileobj( current_file, tempfp )
>>>>>>>   File "/usr/lib/python2.7/dist-packages/duplicity/misc.py", line 170,
>>>>>>> in copyfileobj
>>>>>>>     outfp.write(buf)
>>>>>>> IOError: [Errno 28] No space left on device
>>>>>>>
>>>>>>> The backup chain looks as follows
>>>>>>>
>>>>>>> Chain start time: Thu Aug 16 00:00:27 2012
>>>>>>> Chain end time: Thu Aug 30 00:00:23 2012
>>>>>>> Number of contained backup sets: 12
>>>>>>> Total number of contained volumes: 3477
>>>>>>>  Type of backup set:                            Time:      Num volumes:
>>>>>>>                 Full         Thu Aug 16 00:00:27 2012              2916
>>>>>>>          Incremental         Sun Aug 19 00:00:25 2012                96
>>>>>>>          Incremental         Mon Aug 20 00:00:28 2012                33
>>>>>>>          Incremental         Tue Aug 21 00:00:30 2012                37
>>>>>>>          Incremental         Wed Aug 22 00:00:25 2012                58
>>>>>>>          Incremental         Thu Aug 23 00:00:30 2012                62
>>>>>>>          Incremental         Fri Aug 24 00:00:31 2012                32
>>>>>>>          Incremental         Sat Aug 25 00:00:26 2012                81
>>>>>>>          Incremental         Sun Aug 26 00:00:28 2012                75
>>>>>>>          Incremental         Mon Aug 27 00:00:21 2012                 8
>>>>>>>          Incremental         Tue Aug 28 00:00:18 2012                21
>>>>>>>          Incremental         Thu Aug 30 00:00:23 2012                58
>>>>>>>
>>>>>>> Duplicity is 0.6.18.
>>>>>>>
>>>>>>> Thanks in advance,
>>>>>>>
>>>>>>
>>>>>> you need 30GB (size of file to restore) plus the size of one volume in 
>>>>>> wherever TMP points to.
>>>>>
>>>>> Thanks. But I have only one partition, TMP is unset (if I understand
>>>>> correctly, then it defaults to /tmp), the volume size is default, so
>>>>> I'd need 30GB + 25MB, and I have 80GB free, but apparently that's not
>>>>> enough?
>>>>>
>>>>
>>>> should be... runthe restore with maximum verbosity '-v9' and post the 
>>>> complete output to pastebin (obfuscate private info in it) and send the 
>>>> link. maybe i'll see something.
>>>
>>> I'm attaching the compressed log. It's 2.5MB uncompressed, that's too
>>> big for pastebin. Please let me know if I should it send it in some
>>> other way.
>>>
>>>> you might want to monitor the disk usage during the restore. my guess 
>>>> would be that the final copy of 30GB fails. maybe duplicity keeps 
>>>> downloaded volumes in temp until finished?
>>>
>>> I got the "low disk space, 200MB remaining" warning on the volume
>>> which had 80GB free initially. Looking at the log file, I guess it's
>>> the initial downloading that fails. But why does it have to download
>>> so much?
>>>
>>> I will attach a volume with some 200GB free, point TMP to it, and
>>> restart the restore now.
>>
>> I have retried with TMP pointing to a volume with 244GB and the
>> restore sill fails, although slightly differently. I have attached the
>> compressed log.
>>
> 
> can you verify that during the course of the restore 
>  /media/Sandelys/tmp/duplicity-*-tempdir/
> fills up the containing file system?
> 
> this is suggested by the debug output. trying to pinpoint your issue here.
> 
> 

unlikely but possible.. could you check that 

- you have enough inodes free (df -i) also during the course of the restore
- the file systems are sane by fsck'ing them as a precaution action

..ede