Gluster users - May 2017 - How fill my volume for test?

Hi,
Do you mean to fill your gluster volume?
If so, without getting into the details, I can try giving some pointers:
1)As I see you want to fill data randomly with many directories and files, you
can try to download linux kernel(kernel.org) and untar it
2)there are different tools available in github like FIO,
smallfile(https://github.com/bengland2/smallfile/blob/master/smallfile_cli.py),
crefi
3)One last hack-->If the intention is to fill your volume as fast of possible
and assuming your volume is inside a directory on a thin lv(not directly using
the mount path as the brick ), then you can directly copy  data into the LV
mount with some random data or huge files(any log files or use above tools)
  This helps fill faster, as you are avoiding network as in a mount.
  IMPORTANT NOTE: I would like to reinstate, that you must be filling into the
LV mount instead of the brick mount. Eg: if you LV mount is /rhs/lv1, I
expect(as recommended), your brick path is something like
/rhs/lv1/<brickname> . Hence you can fill in data under
/rhs/lv1/<datadump>


Regards,
nag

----- Original Message -----
From: "Tahereh Fattahi" <t28.fattahi at gmail.com>
To: gluster-users at gluster.org
Sent: Tuesday, 30 May, 2017 3:29:40 PM
Subject: [Gluster-users] How fill my volume for test?

Hi 
I need to fill my volume and after that start testing. 
Is there a test case, test bench (I dont know the name) a program that create a
lot of directory and files with different type and size (the data in files is
not important, maybe random)?
I can do this by my self but I dont know is this work correct? 
Is there any standard for fill a file system? 

_______________________________________________
Gluster-users mailing list
Gluster-users at gluster.org
http://lists.gluster.org/mailman/listinfo/gluster-users

Thank you
Is it correct that there is not a standard way to fill a file system?
If I fill my gluster volume with each of these methods, the result of
performance test (speed of read, write, ...) are valid?

On Tue, May 30, 2017 at 4:22 PM, Nag Pavan Chilakam <nchilaka at
redhat.com>
wrote:
> Hi,
> Do you mean to fill your gluster volume?
> If so, without getting into the details, I can try giving some pointers:
> 1)As I see you want to fill data randomly with many directories and files,
> you can try to download linux kernel(kernel.org) and untar it
> 2)there are different tools available in github like FIO, smallfile(
> https://github.com/bengland2/smallfile/blob/master/smallfile_cli.py),
> crefi
> 3)One last hack-->If the intention is to fill your volume as fast of
> possible and assuming your volume is inside a directory on a thin lv(not
> directly using the mount path as the brick ), then you can directly copy
> data into the LV mount with some random data or huge files(any log files or
> use above tools)
>   This helps fill faster, as you are avoiding network as in a mount.
>   IMPORTANT NOTE: I would like to reinstate, that you must be filling into
> the LV mount instead of the brick mount. Eg: if you LV mount is /rhs/lv1, I
> expect(as recommended), your brick path is something like
> /rhs/lv1/<brickname> . Hence you can fill in data under
/rhs/lv1/<datadump>
>
>
> Regards,
> nag
>
> ----- Original Message -----
> From: "Tahereh Fattahi" <t28.fattahi at gmail.com>
> To: gluster-users at gluster.org
> Sent: Tuesday, 30 May, 2017 3:29:40 PM
> Subject: [Gluster-users] How fill my volume for test?
>
> Hi
> I need to fill my volume and after that start testing.
> Is there a test case, test bench (I dont know the name) a program that
> create a lot of directory and files with different type and size (the data
> in files is not important, maybe random)?
> I can do this by my self but I dont know is this work correct?
> Is there any standard for fill a file system?
>
> _______________________________________________
> Gluster-users mailing list
> Gluster-users at gluster.org
> http://lists.gluster.org/mailman/listinfo/gluster-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.gluster.org/pipermail/gluster-users/attachments/20170531/5b536acf/attachment.html>