[Gluster-users] nginx using glusterfs volume issue, can't read concurrently

[符永涛]

Seems a performance issue with nginx serving large file. I configure nginx
with the following options and it's much better now. I'll make more test
and get back updated.

sendfile        off;

output_buffers 1 1024k;

aio on;

directio 1024;

2012/8/18 符永涛 <yongtaofu at gmail.com>

> Can somebody provide some insight on this issue? Thank you very much.
>
>
> 2012/8/17 符永涛 <yongtaofu at gmail.com>
>
>> Sorry to tell you it is not related to ubuntu, nginx not working on
>> glusterfs volume both on redhat and ubuntu.
>>
>>
>> 2012/8/17 符永涛 <yongtaofu at gmail.com>
>>
>>> It occurs on redhat 6.3 with kernel 2.6.32-279.el6.x86_64, but nginx
>>> works fine on glusterfs volume on ubuntu 12.
>>> Any suggestions?
>>>
>>>
>>> 2012/8/17 符永涛 <yongtaofu at gmail.com>
>>>
>>>> Dear experts,
>>>>
>>>> Currently I encountered an issue about nginx using glusterfs volume as
>>>> webroot.
>>>> It ends up with nginx can't read file concurrently.
>>>>
>>>> for example if I download an large file from nginx http server from
>>>> two clients one of them will block.
>>>>
>>>> for example
>>>> [root at 10.10.135.21]$ curl
>>>> 10.10.135.21:45678/mnt/teamshare/software/testfolder/test.10gfile -v
>>>> > /dev/null
>>>>
>>>> [root at 10.10.135.23]$ curl
>>>> 10.10.135.21:45678/mnt/teamshare/software/testfolder/test.10gfile -v
>>>> > /dev/null
>>>>
>>>> the second download will block.
>>>>
>>>> The issue only occur when nginx work with glusterfs volume. Apache with
>>>> glusterfs volume is fine.
>>>> My glusterfs version is 3.2.5.
>>>>
>>>> Any one known what's the root cause?
>>>> Thank you very much!
>>>>
>>>> --
>>>> 符永涛
>>>>
>>>
>>>
>>>
>>> --
>>> 符永涛
>>>
>>
>>
>>
>> --
>> 符永涛
>>
>
>
>
> --
> 符永涛
>



-- 
符永涛
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-users/attachments/20120818/853b25e4/attachment.html>