As a part of my initial contribution to the organization, I have done the
1. Installed gluster on 2 nodes on digitalocean
2. Specs of the nodes: 2 GB Memory / 40 GB Disk / NYC2 - Ubuntu 16.04.1
3. Set up a volume and mounted it
4. Set up a client machine, and ran the benchmark tests from there only.
After the setup of the nodes and the client machine, I ran the benchmark
tests from . I found the following issues while creating the benchmark:
- there were no instructions on the latest ppa for gluster, so I had to
search on launchpad for Ubuntu
- mount_point is a required variable which is not accepted properly
- no instructions on setting up IOZone.
- IOZone configuration file isn't generated properly for IOZone3.
- smallfile is configured on all clients, however the check for
smallfile is done on localhost, so even though it's installed it installs
- error handling is weak on tests
- the mounted volume name is hard coded for smallfile
I would like to improve the benchmark script. Also, could I make a pull
request for the same?
On fixing the errors and after some research I was able to generate the
following results for the benchmark:
Sequential Writes 64 record size: 549,826 kBps
Sequential Reads 64 record size: 919,270 kBps
Random Writes 64 record size: 278,564 kBps Random Reads 64 record size:
81,914 KBps Smallfile Creates 64 file size: 236 files/sec Smallfile Reads
64 file size: 627 files/sec
Smallfile ls -l 64 file size: 10693 files/sec
On 14 October 2016 at 21:27, Shyam <srang...@redhat.com> wrote:
> On 10/14/2016 10:48 AM, Manikandan Selvaganesh wrote:
>> Hi Soumya,
>> Welcome to the community.
>> Here is the link for Gluster Documentation. I would suggest you to
>> google and
>> read a bit about GlusterFS and then get started with "Quick Start
>> Once you have done your setup and have played a bit around the
>> installation and
>> configuration move on with "Developers Guide".
>> If you want to get started with Code contributions pick some EasyFix
>> bugs which
>> can be found here. After this I hope you would have got a minimal
>> idea and then
>> explore more in depth and pick up the project/component which interests
>> you more.
>> Again, we have some list of projects already listed, check out if
>> anything interests
>> you here. Feel free to bring your own ideas as well. These are quite
>> generic for anyone
>> who is new to the community and in case if you want to know
>> specifically about
>> Outreachy, someone in the community will surely respond to you shortly.
> Let me take the Outreachy part up.
> There are 2 projects there, one relating to the documentation, for which
> Manikandan has filled in some links and thoughts. The other being the
> instrumentation tooling around performance.
> For the latter, I would suggest that you get a gluster volume up and
> running, and attempt the GlusterBench.py  against it, and start with
> reporting the results. Again, Manikandan has covered getting gluster up and
> running. For any questions, or things that you get stuck on when running
> the bench script, post back here and we will help as needed.
>> If you have queries, please mail us back. Also, we are always available
>> on #gluster-dev
>> and #gluster-meeting in Freenode.
>> All the best :-)
>>  https://gluster.readthedocs.io/en/latest/
>>  https://gluster.readthedocs.io/en/latest/Quick-Start-Guide/
>>  https://gluster.readthedocs.io/en/latest/Developer-guide/Dev
>>  https://gluster.readthedocs.io/en/latest/Developer-guide/Eas
>>  https://gluster.readthedocs.io/en/latest/Developer-guide/Projects/
>  GlusterBench.py : https://github.com/gluster/gbe
>> Manikandan Selvaganesh.
>> On Fri, Oct 14, 2016 at 7:56 PM, Ms ms <ms.soumy...@gmail.com
>> <mailto:ms.soumy...@gmail.com>> wrote:
>> I'm a research student pursuing my Masters in IIIT-Hyderabad. I am
>> keen on working on Gluster's Outreachy project.
>> I have prior experience in configuring, maintaining and managing
>> systems in an MHRD project. I have completed the required course
>> credits towards my degree and am working on my Thesis currently. It
>> would be great opportunity for me to learn and contribute to the
>> project as well.
>> As I am a bit new to the community it would be nice if anyone can
>> guide me a few useful resources to get me started.
>> Thanks and regards,
>> Gluster-users mailing list
>> Glusterfirstname.lastname@example.org <mailto:Glusteremail@example.com>
>> Gluster-users mailing list
Gluster-users mailing list