Re: Lower IO performance with 3.7.x than with 3.5.x

[Krutika Dhananjay <kdhananj@xxxxxxxxxx>]

CC'ing distribute component owners.

-Krutika

---

From: "Geoffrey Letessier" <geoffrey.letessier@xxxxxxx>
To: gluster-users@xxxxxxxxxxx
Sent: Thursday, August 13, 2015 5:39:47 PM
Subject: [Gluster-users] Lower IO performance with 3.7.x than with 3.5.x

Hi,

Previously we used GlusterFS v3.5.3 and we could achieve around 1.1GBs for read/write in plain distributed volume. 

Now, since I have upgraded GlusterFS to 3.7.3, i can only achieve around 700MBs (max).

Here some additional information concerning my volume. 

# gluster volume info vol_workdir_amd

 

Volume Name: vol_workdir_amd

Type: Distribute

Volume ID: 42d8f9fa-3b24-414f-a275-dadcfdd8d3db

Status: Started

Number of Bricks: 4

Transport-type: tcp,rdma

Bricks:

Brick1: ib-storage1:/export/brick_workdir/brick1/data

Brick2: ib-storage3:/export/brick_workdir/brick1/data

Brick3: ib-storage1:/export/brick_workdir/brick2/data

Brick4: ib-storage3:/export/brick_workdir/brick2/data

Options Reconfigured:

client.event-threads: 4

performance.readdir-ahead: on

auth.allow: 10.0.*

cluster.min-free-disk: 5%

performance.cache-size: 1GB

performance.io-thread-count: 32

diagnostics.brick-log-level: CRITICAL

nfs.disable: on

performance.read-ahead: on

and the status is OK for all bricks in.

Thanks in advance,

Geoffrey

------------------------------------------------------
Geoffrey Letessier
Responsable informatique & ingénieur système
UPR 9080 - CNRS - Laboratoire de Biochimie Théorique
Institut de Biologie Physico-Chimique
13, rue Pierre et Marie Curie - 75005 Paris
Tel: 01 58 41 50 93 - eMail: geoffrey.letessier@xxxxxxx

_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users

_______________________________________________
Gluster-users mailing list
Gluster-users@xxxxxxxxxxx
http://www.gluster.org/mailman/listinfo/gluster-users