Bug 1232165 - NFS Authentication Performance Issue
Summary: NFS Authentication Performance Issue
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: GlusterFS
Classification: Community
Component: nfs
Version: mainline
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
Assignee: Niels de Vos
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks: 1231366
TreeView+ depends on / blocked
 
Reported: 2015-06-16 08:20 UTC by Niels de Vos
Modified: 2016-06-16 13:12 UTC (History)
1 user (show)

Fixed In Version: glusterfs-3.8rc2
Clone Of:
Environment:
Last Closed: 2016-06-16 13:12:06 UTC
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Embargoed:


Attachments (Terms of Use)

Description Niels de Vos 2015-06-16 08:20:54 UTC
+++ This bug was initially created as a clone of Bug #1231366 +++

Description of problem:
When file operations are sent to the NFS server, authorized filehandles are cached
using the exportid, mountid, gfid and host as the key to the cache. This meant that any file OR directory will always fail on the *first* fop to that filehandle since the
cache used the gfid as part of the key to the cache. However, if an export is authorized, this effectively means that ALL subdirectories and files in the export directory are authorized per the permissions of the export. This results slow times to walking a directory structure over an NFS mount.

Version-Release number of selected component (if applicable): 3.7.1


How reproducible: Turn on authentication, make a large netgroups file and walk the directory


Steps to Reproduce:
1. Make a large netgroups file, create an export and assign the netgroup to that export
2. GDB and break on is_nfs_fh_cached, and you'll see many cache misses
3. Walk the filesystem and notice slower than expected performance

Actual results: Many misses in the auth cache, resulting in slower walks of the filesystem


Expected results: More hits in the auth cache, resulting in faster walks of the filesystem.


Additional info:

Comment 1 Anand Avati 2015-06-16 08:32:25 UTC
REVIEW: http://review.gluster.org/11245 (nfs: Authentication performance improvements) posted (#1) for review on master by Niels de Vos (ndevos)

Comment 2 Anand Avati 2015-06-16 12:15:15 UTC
COMMIT: http://review.gluster.org/11245 committed in master by Kaleb KEITHLEY (kkeithle) 
------
commit a9f58cd6cfc29a47af868fff29ce3133b9f9efe5
Author: Shreyas Siravara <sshreyas>
Date:   Fri Jun 12 13:11:35 2015 -0700

    nfs: Authentication performance improvements
    
    When file operations are sent to the NFS server, authorized filehandles
    are cached using the exportid, mountid, gfid and host as the key to the
    cache. This meant that any file OR directory will always fail on the
    *first* fop to that filehandle since the cache used the gfid as part of
    the key to the cache. However, if an export is authorized, this
    effectively means that ALL subdirectories and files in the export
    directory are authorized per the permissions of the export. This results
    slow times to walking a directory structure over an NFS mount.
    
    Change-Id: Iad811ad7255b454d1712e75a637478401d40791e
    BUG: 1232165
    Signed-off-by: Shreyas Siravara <sshreyas>
    Signed-off-by: Niels de Vos <ndevos>
    Reviewed-on: http://review.gluster.org/11245
    Reviewed-by: jiffin tony Thottan <jthottan>
    Reviewed-by: Kaleb KEITHLEY <kkeithle>
    Tested-by: Gluster Build System <jenkins.com>

Comment 3 Niels de Vos 2016-06-16 13:12:06 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-3.8.0, please open a new bug report.

glusterfs-3.8.0 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] http://blog.gluster.org/2016/06/glusterfs-3-8-released/
[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.