[Bugs] [Bug 1535852] glusterfind is extremely slow if there are lots of changes

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1535852

Vinayak Papnoi <vpapnoi at redhat.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ON_QA                       |VERIFIED



--- Comment #6 from Vinayak Papnoi <vpapnoi at redhat.com> ---
Tested the glusterfind run on 3.4.0 vs 3.3.1 for the same volume topology and
same I/O operation (in this case linux kernel untar, approximately 1.9GB post
untar).

Following are my observations:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Build : glusterfs-3.8.4-54.6.el7rhgs.x86_64

[root at dhcp42-39 ~]# gluster v info

Volume Name: alpha
Type: Distributed-Replicate
Volume ID: cfea6bcc-13ec-4305-b580-6f8e4a391cf8
Status: Started
Snapshot Count: 0
Number of Bricks: 2 x 3 = 6
Transport-type: tcp
Bricks:
Brick1: 10.70.42.39:/bricks/brick0/alpha-b1
Brick2: 10.70.42.22:/bricks/brick0/alpha-b2
Brick3: 10.70.42.21:/bricks/brick0/alpha-b3
Brick4: 10.70.42.14:/bricks/brick0/alpha-b4
Brick5: 10.70.42.15:/bricks/brick0/alpha-b5
Brick6: 10.70.42.16:/bricks/brick0/alpha-b6
Options Reconfigured:
changelog.capture-del-path: on
changelog.changelog: on
storage.build-pgfid: on
transport.address-family: inet
nfs.disable: on
cluster.enable-shared-storage: enable
cluster.brick-multiplex: enable
cluster.max-bricks-per-process: 20
nfs-ganesha: enable

[root at dhcp42-39 scripts]# time glusterfind pre session1 alpha
/tmp/outfilenew.txt --regenerate-outfile
Generated output file /tmp/outfilenew.txt

real    7m14.345s
user    4m54.421s
sys    2m2.892s

-------------------------------


Build : glusterfs-3.12.2-8.el7rhgs.x86_64

[root at dhcp43-18 ~]# gluster v info

Volume Name: alpha
Type: Distributed-Replicate
Volume ID: 925b3b99-4742-4ecd-b462-acb6aec1d1a9
Status: Started
Snapshot Count: 0
Number of Bricks: 2 x 3 = 6
Transport-type: tcp
Bricks:
Brick1: 10.70.43.18:/bricks/brick0/alpha-b1
Brick2: 10.70.43.9:/bricks/brick0/alpha-b2
Brick3: 10.70.43.2:/bricks/brick0/alpha-b3
Brick4: 10.70.42.240:/bricks/brick0/alpha-b4
Brick5: 10.70.43.4:/bricks/brick0/alpha-b5
Brick6: 10.70.42.243:/bricks/brick0/alpha-b6
Options Reconfigured:
changelog.capture-del-path: on
changelog.changelog: on
storage.build-pgfid: on
nfs.disable: on
snap-activate-on-create: disable
auto-delete: disable
cluster.brick-multiplex: enable
cluster.enable-shared-storage: disable

[root at dhcp43-18 ~]# time glusterfind pre session1 alpha /tmp/outfilenew.txt
Generated output file /tmp/outfilenew.txt

real    0m7.588s
user    0m1.751s
sys     0m0.998s

-------------------------------

As seen above, the 3.4.0 build was able to run glusterfind about 60 times
faster than 3.3.1.

Hence, moving bug to VERIFIED.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug https://bugzilla.redhat.com/token.cgi?t=vJkbOi1qll&a=cc_unsubscribe