[Bugs] [Bug 1200457] gluster performance issue as data is added to volume. tar extraction of files goes from 1-minute on empty volume to 20-minutes on volume with 40TB.

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1200457



--- Comment #18 from Manoj Pillai <mpillai at redhat.com> ---
(In reply to Manoj Pillai from comment #17)

More details on the test:

Versions:
glusterfs*-3.8.2-1.el7.x86_64
RHEL 7.2 (server): kernel-3.10.0-327.el7.x86_64
RHEL 7.1 (clients): kernel-3.10.0-229.el7.x86_64

Benchmarking Tool: smallfile version 3.0
[https://github.com/bengland2/smallfile]


smallfile_cli.py --top ${top_dir} --host-set ${hosts_str} --threads 4 --files
8192 --file-size 64 --record-size 64 --fsync Y --response-times N --operation
create

There are 4 clients, so no. of files in each iteration is: (4 clients) * (4
threads per client) * 8192 = 128K files. file size is 64KB.

Each iteration creates a new "top directory" under the mountpoint and all files
created in that iteration are created inside that sub-directory in a hierarchy
with 100 files per directory (default).

Gluster volume info:
Volume Name: perfvol
Type: Distribute
<...>
Status: Started
Number of Bricks: 1
Transport-type: tcp
Bricks:
<...>
Options Reconfigured:
cluster.lookup-optimize: on
server.event-threads: 4
client.event-threads: 8
performance.client-io-threads: on
transport.address-family: inet
performance.readdir-ahead: on

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.