0

I tar ~/Documents to create a backup and my .tgz file turns out to be about 683 MB. I allow my Ubuntu One upload speed to be about 2.5 Mb/s. The sync takes several hours. My Power Management setting is "never suspend". I might on some occasions reboot Ubuntu in the middle of the sync in order to switch to Windows to print a file since Ubuntu does not have a driver for my printer. I get no complaints so this seems like a permissible operation. After the sync completed I downloaded the file from the Ubuntu One website. Strangely the file is now 619 MB and using the diff utility I can see a message that says the binary files are not identical.

Double-clicking the downloaded .tgz file opens the Archive Manager and I get this:

gzip: stdin: unexpected end of file
tar: Unexpected EOF in archive
tar: Error is not recoverable: exiting now

How did this file get corrupted?

2
  • Does just doing "tar xvf file.tgz" work? If you downloaded from the web site, it is possible that your browser saved the file un-gzipped already.
    – dobey
    Jun 3, 2012 at 0:40
  • It will extract via "tar xvf file.tgz" but then at some point it will fail with the same message as provided by Archive Manager.
    – H2ONaCl
    Jun 4, 2012 at 10:32

2 Answers 2

0

I might on some occasions reboot Ubuntu in the middle of the sync

I think this will be the answer. Also, uploading a big file instead of a bunch of small files is more dangerous. I think you should drop the creating a compressed archive part, and just set Ubuntu One to synchronize your ~/Documents folder.

Alternative solution

I have read an issue in the comments to this question that the files should be revised between two occasions (version control). One way would be setting up a version control system, but that is a big amount of hassle for home usage.

So, if you wish to keep revisions of files, but to use a solution that does not compress the files, you can have a small file to create a copy of your Documents folder with an added timestamp, and let that folder (as the folder uncompressed) synced to Ubuntu One.

But if the size of the folder is big, this download causes too much write on the system (not very good for SSD-based storages) and increases latency. I'm not sure how Ubuntu One treates symbolical links, though... If symbolical links are treated as copies of the file (so the original content is uploaded, not the link inode), you can just create symlinks.

5
  • It seems absurd if Ubuntu One does not CRC the file on both ends and check for a match before reporting the sync is complete.
    – H2ONaCl
    Jun 2, 2012 at 12:21
  • While it does sound absurd, it might be the case. I have upadted the answer with an alternative solution to match your needs described in your comment to izx's post.
    – Whisperity
    Jun 2, 2012 at 12:34
  • Thanks for the suggestion. Without compression I would be bothered by the inefficiency. Unfortunately .tgz also seems inefficient because I am seeing only 2:1 compression. I might have to post on superuser to get an alternative cloud recommendation for .tgz backups.
    – H2ONaCl
    Jun 2, 2012 at 13:14
  • @broiyan Unless all the files in your ~/Documents folder are plain text, you are probably going to see somewhat inefficient compression. Compressing already compressed files isn't very efficient. There's no reason the file should be corrupt when uploading to Ubuntu One though.
    – dobey
    Jun 3, 2012 at 0:39
  • @dobey, Without the reboot, the sync and download resulted in a successful binary comparison.
    – H2ONaCl
    Jun 3, 2012 at 5:19
-1

I have had problems with 5-6 MB zip files being corrupted by Ubuntu One. I create a ZIP in Windows, add them to an Ubuntu One folder in Linux and download the zip on another Windows machine. Corrupt file.

So, instead, I copy the zip from the Ubuntu One folder to a USB stick and open the the zip on the second Windows machine and it's OK. So the difference is the upload to Ubuntu One (via the Ubuntu One software in Ubuntu) and the download from the Ubuntu One website, using Internet Explorer.

The problem has happened two times to me, with similar files involved.

In your case, does the web site report the size 619 MB or 683 MB? This could indicate if the problem is caused during upload or download. If it reports 683 MB, then maybe it's worth trying a different browser to download with.

1
  • Short answer - use Firefox to download the file. I uploaded a 6 549 092 bytes zip file, which is 6.2 MiB, using the Ubuntu application. The web lists the file as 6.2 MB. Downloading the file with Firefox in Windows reports a size of 6.12 MB (incorrect) during download, but successfully downloads all 6 549 092 bytes. Downloading the file with IE 9 in Windows fails on the first attempt, but succeeds on second try. Ubuntu One apparently reports an incorrect file size during download.
    – Christer
    Mar 31, 2013 at 12:57

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .