copy big file (2GB) to encrypted filesystem through samba share

Bug #202200 reported by wy
10
This bug affects 1 person
Affects Status Importance Assigned to Milestone
EncFS
Fix Released
Wishlist
rosa maria

Bug Description

I used debian 4.0 and made a samba public r/w share so that windows can access it. even I use latest encfs 1.4.1.1. this problem is still here. here is log by enabling encfs verbose option.

command:
encfs --public -fv /home/encrypted /home/plain.

/home/plain is shared by samba server.

It kept printing this message, after 10 minutes, it gave up.

09:09:51 (BlockFileIO.cpp:340) padding block 4260841
09:09:51 (BlockFileIO.cpp:340) padding block 4260842
09:09:51 (BlockFileIO.cpp:340) padding block 4260843
09:09:51 (BlockFileIO.cpp:340) padding block 4260844
09:09:51 (BlockFileIO.cpp:340) padding block 4260845
09:09:51 (BlockFileIO.cpp:340) padding block 4260846
09:09:51 (BlockFileIO.cpp:340) padding block 4260847
09:09:51 (BlockFileIO.cpp:340) padding block 4260848
09:09:51 (BlockFileIO.cpp:340) padding block 4260849
09:09:51 (BlockFileIO.cpp:340) padding block 4260850
09:09:51 (BlockFileIO.cpp:340) padding block 4260851
09:09:51 (BlockFileIO.cpp:340) padding block 4260852
09:09:51 (BlockFileIO.cpp:340) padding block 4260853
09:09:51 (BlockFileIO.cpp:340) padding block 4260854
09:09:51 (BlockFileIO.cpp:340) padding block 4260855
09:09:51 (BlockFileIO.cpp:340) padding block 4260856
09:09:51 (BlockFileIO.cpp:340) padding block 4260857
09:09:51 (BlockFileIO.cpp:340) padding block 4260858
09:09:51 (BlockFileIO.cpp:340) padding block 4260859
09:09:51 (BlockFileIO.cpp:340) padding block 4260860
09:09:51 (BlockFileIO.cpp:340) padding block 4260861
09:09:51 (BlockFileIO.cpp:340) padding block 4260862
09:09:51 (BlockFileIO.cpp:340) padding block 4260863
09:09:51 (BlockFileIO.cpp:340) padding block 4260864
09:09:51 (BlockFileIO.cpp:340) padding block 4260865
09:09:51 (BlockFileIO.cpp:340) padding block 4260866
09:09:51 (DirNode.cpp:805) in FileNodeDestructor for /home/encrypted/7e8SVj5,kOlkMzwj0hsHNwip (refcount 1)
09:09:51 (encfs.cpp:79) getattr on /home/encrypted/7e8SVj5,kOlkMzwj0hsHNwip
09:09:51 (DirNode.cpp:805) in FileNodeDestructor for /home/encrypted/7e8SVj5,kOlkMzwj0hsHNwip (refcount 1)
09:09:52 (DirNode.cpp:805) in FileNodeDestructor for /home/encrypted/se5i49psFQ7Gh7FzMUuImMqi (refcount 1)
09:09:52 (DirNode.cpp:815) destroying FNode /home/encrypted/se5i49psFQ7Gh7FzMUuImMqi
09:09:52 (DirNode.cpp:805) in FileNodeDestructor for /home/encrypted/ (refcount 1)
09:09:52 (DirNode.cpp:815) destroying FNode /home/encrypted/
09:09:52 (DirNode.cpp:309) Leaked FileNodes: 1
09:09:52 (DirNode.cpp:313) Leaked node: /home/encrypted/7e8SVj5,kOlkMzwj0hsHNwip

You can see error message at windows side.

Cannot copy xp400004: the specified network name is no longer available

Revision history for this message
wy (wmyang2000) wrote :

I found the root problem. It is in BlockFileIO::padFile. system issues truncate command. it takes long time to do this because it is a big file and connection times out.

what is the best way to solve this problem?

thanks,

Revision history for this message
Valient Gough (vgough) wrote :

Attached is a patch that will make encfs leave holes in files. If a block is read as all zeros, it will be assumed to be a hole and will be left as 0's when read (not deciphered).

Give that a try. If that works well for you, I'll look at putting that into encfs as an option.

Changed in encfs:
importance: Undecided → Wishlist
Revision history for this message
wy (wmyang2000) wrote :

thank you. it works.

Revision history for this message
Valient Gough (vgough) wrote :

I've added an option to enable file holes. It will be part of the next release.

Changed in encfs:
status: New → Fix Committed
Valient Gough (vgough)
Changed in encfs:
status: Fix Committed → Fix Released
Revision history for this message
Webu (webu) wrote :

Thank you for doing that, I had the same problem transferring a 8GB file from Windows XP computer to Ubuntu server 8.04 running Samba + EncFS on top of ext3 shares.

I fought with the problem for many hours (without any success) until I magically found this Lauchpad, compiled the latest release and reconstructed all my encrypted folders to use the new feature. After that everything worked perfectly, like I was using an unencrypted Samba share :-) The 8GB file just flew through my network and also with medium sized files there was no wait (pauses) anymore between different files or when starting the transfer (it took a while for EncFS to "reserve" the disk space of upcoming file(s) before the actual transfer could begin).

Wehee!

Changed in encfs:
assignee: nobody → rosa maria (rprosamaria383)
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.