Bug 1317863 - glfs_dup() functionality is broken
Summary: glfs_dup() functionality is broken
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: GlusterFS
Classification: Community
Component: libgfapi
Version: 3.7.9
Hardware: All
OS: All
low
medium
Target Milestone: ---
Assignee: bugs@gluster.org
QA Contact: Sudhir D
URL:
Whiteboard:
Depends On: 1311146
Blocks:
TreeView+ depends on / blocked
 
Reported: 2016-03-15 11:49 UTC by rjoseph
Modified: 2016-04-19 06:59 UTC (History)
6 users (show)

Fixed In Version: glusterfs-3.7.10
Doc Type: Bug Fix
Doc Text:
Clone Of: 1311146
Environment:
Last Closed: 2016-04-19 06:59:50 UTC
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Embargoed:


Attachments (Terms of Use)

Description rjoseph 2016-03-15 11:49:39 UTC
+++ This bug was initially created as a clone of Bug #1311146 +++

Description of problem:
***********************
glfs_dup() functionality is currently broken in glusterfs.


Steps to Reproduce: 
*******************
Code snippet from libgfapi-python tests:

    f = File(self.vol.open("dune", os.O_CREAT | os.O_EXCL | os.O_RDWR))
    f.write("I must not fear. Fear is the mind-killer.")
    fdup = f.dup()
    self.assertTrue(isinstance(fdup, File))
    f.close()
    ret = fdup.lseek(0, os.SEEK_SET)
    self.assertEqual(ret, 0)


Actual results:
***************
ERROR: test_write_file_dup_lseek_read (test.functional.libgfapi-python-tests.FileOpsTest)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/root/src/libgfapi-python/test/functional/libgfapi-python-tests.py", line 236, in test_write_file_dup_lseek_read
    ret = fdup.lseek(0, os.SEEK_SET)
  File "/root/src/libgfapi-python/gluster/gfapi.py", line 306, in lseek
    raise OSError(err, os.strerror(err))
OSError: [Errno 9] Bad file descriptor


Expected results:
*****************
lseek() should succeed.


Additional information
**********************
This change (http://review.gluster.org/13340) disallows any operation on duplicate once the original fd is closed.

Seems like a deep copy of original fd instance should be done.

--- Additional comment from Vijay Bellur on 2016-03-10 05:14:36 EST ---

REVIEW: http://review.gluster.org/13666 (gfapi: preserve glfd state during glfs_dup) posted (#1) for review on master by Rajesh Joseph (rjoseph)

--- Additional comment from  on 2016-03-10 05:59:58 EST ---

glfs_dup is completely broken. Current behavior is not in accordance with POSIX dup (http://linux.die.net/man/2/dup)

http://review.gluster.org/13666 only addresses are very small subset of the problem. glfs_dup needs to be re-written to fix the broken behavior. Therefore not changing the state of the bug as of now.

--- Additional comment from Vijay Bellur on 2016-03-15 02:46:39 EDT ---

COMMIT: http://review.gluster.org/13666 committed in master by Niels de Vos (ndevos) 
------
commit 5bdfaf98904a339144bf3a237b162e8385b95085
Author: Rajesh Joseph <rjoseph>
Date:   Wed Mar 9 22:05:18 2016 +0530

    gfapi: preserve glfd state during glfs_dup
    
    Following patch introduced a new state variable in glfd
    to track the current status of the fd.
    http://review.gluster.org/13340/
    
    But this state was not copied in glfd_dup function.
    
    BUG: 1311146
    Change-Id: I283f8944035f6defe491f81e13d7ef28fc440572
    Signed-off-by: Rajesh Joseph <rjoseph>
    Reviewed-on: http://review.gluster.org/13666
    Smoke: Gluster Build System <jenkins.com>
    Reviewed-by: Prashanth Pai <ppai>
    Tested-by: Prashanth Pai <ppai>
    NetBSD-regression: NetBSD Build System <jenkins.org>
    CentOS-regression: Gluster Build System <jenkins.com>
    Reviewed-by: Raghavendra Talur <rtalur>
    Reviewed-by: Niels de Vos <ndevos>

Comment 1 Vijay Bellur 2016-03-15 11:55:52 UTC
REVIEW: http://review.gluster.org/13742 (gfapi: preserve glfd state during glfs_dup) posted (#1) for review on release-3.7 by Rajesh Joseph (rjoseph)

Comment 2 Vijay Bellur 2016-03-29 10:00:13 UTC
REVIEW: http://review.gluster.org/13742 (gfapi: preserve glfd state during glfs_dup) posted (#2) for review on release-3.7 by Rajesh Joseph (rjoseph)

Comment 3 Vijay Bellur 2016-03-30 14:16:29 UTC
REVIEW: http://review.gluster.org/13742 (gfapi: preserve glfd state during glfs_dup) posted (#3) for review on release-3.7 by Niels de Vos (ndevos)

Comment 4 Vijay Bellur 2016-03-30 14:23:36 UTC
COMMIT: http://review.gluster.org/13742 committed in release-3.7 by Niels de Vos (ndevos) 
------
commit 4b9853358a8fbec0876383e6b7839e39cc025c0c
Author: Rajesh Joseph <rjoseph>
Date:   Wed Mar 9 22:05:18 2016 +0530

    gfapi: preserve glfd state during glfs_dup
    
    Following patch introduced a new state variable in glfd
    to track the current status of the fd.
    http://review.gluster.org/13340/
    
    But this state was not copied in glfd_dup function.
    
    Backport of commit 5bdfaf98904a339144bf3a237b162e8385b95085:
    > BUG: 1311146
    > Change-Id: I283f8944035f6defe491f81e13d7ef28fc440572
    > Signed-off-by: Rajesh Joseph <rjoseph>
    > Reviewed-on: http://review.gluster.org/13666
    > Smoke: Gluster Build System <jenkins.com>
    > Reviewed-by: Prashanth Pai <ppai>
    > Tested-by: Prashanth Pai <ppai>
    > NetBSD-regression: NetBSD Build System <jenkins.org>
    > CentOS-regression: Gluster Build System <jenkins.com>
    > Reviewed-by: Raghavendra Talur <rtalur>
    > Reviewed-by: Niels de Vos <ndevos>
    
    Change-Id: I283f8944035f6defe491f81e13d7ef28fc440572
    BUG: 1317863
    Signed-off-by: Rajesh Joseph <rjoseph>
    Reviewed-on: http://review.gluster.org/13742
    NetBSD-regression: NetBSD Build System <jenkins.org>
    CentOS-regression: Gluster Build System <jenkins.com>
    Reviewed-by: Niels de Vos <ndevos>
    Smoke: Gluster Build System <jenkins.com>

Comment 5 Niels de Vos 2016-03-30 14:31:00 UTC
The merged patch only reverts to the previous (broken) behaviour of glfs_dup(). It is not the final fix.

(Added the Tracking keyword to prevent scripts from complaining about the status)

Comment 6 Kaushal 2016-04-19 06:59:50 UTC
This bug is getting closed because a release has been made available that should address the reported issue. In case the problem is still not fixed with glusterfs-3.7.10, please open a new bug report.

glusterfs-3.7.10 has been announced on the Gluster mailinglists [1], packages for several distributions should become available in the near future. Keep an eye on the Gluster Users mailinglist [2] and the update infrastructure for your distribution.

[1] https://www.gluster.org/pipermail/gluster-users/2016-April/026164.html
[2] http://thread.gmane.org/gmane.comp.file-systems.gluster.user


Note You need to log in before you can comment on or make changes to this bug.