[Openstack] [Keystone] Keystone performance work

Neependra Khare nkhare at redhat.com
Tue Dec 17 09:51:24 UTC 2013


Hi Wanghong,

On 12/17/2013 03:10 PM, 王宏 wrote:
> Hi Neependra.
>
> We have done some keystone performance works already at HuaWei. We are 
> interested in the number of PKI_token_authentication or
> PKI_token_revocation_list_get requests that keystone V3 can handle per 
> second.
> We also made some attempts to improve the concurrency value per second.
>
> We had measured the concurrency value against a single-process 
> Pthon-only keystone server, a multi-processes Keystone server running 
> inside an Apache container server and 2, 3, 4 multi-processes Keystone 
> servers running inside an Apache container server.
>
> Could I modify your wiki to share the detail data?
Sure.

Regards,
Neependra

>
> Regards
> wanghong
>
>
> 2013/12/17 Neependra Khare <nkhare at redhat.com <mailto:nkhare at redhat.com>>
>
>     Hi Tristan,
>
>     On 12/16/2013 09:51 PM, Tristan De Cacqueray wrote:
>>     On Wed, 11 Dec 2013 14:41:15 -0500
>>     ayoung atredhat.com  <http://redhat.com>  (Adam Young) wrote:
>>
>>>     Sounds good for a start.
>>>
>>>     There is a lot in Devstack that should help as far as documenting how
>>>     to set up LDAP etc.
>>>
>>>     Can you indicate which is going to be your first effort?  We
>>>     (Keystone team) can provide some guidance on how to best hammer on it.
>>>
>>>
>>     Hello!
>>
>>     So I have started the UUID vs PKI and SQL vs LDAP comparison into one
>>     benchmark using Devstack. I mostly focus on the time it takes to
>>     validate token in those four scenarios, though I measure the time it
>>     takes to create users and generate tokens as well.
>     Thanks for taking it up.
>
>>     I edited the wiki page to document this test, including the scripts
>>     being used. Maybe those should be removed as the page grows...
>>     The results should be ready in a couple of hours.
>     May be we can put the scripts used here on github or somewhere else.
>     Any suggestions ?
>
>>     Feedback are most welcome, and feel free to use/improve the code I
>>     pasted!
>     - I think we need to test with more numbers of users.
>     - We should consider scenarios when creation, generation and
>     validation happening in parallel.
>     - We should do the test when there is already some existing
>     dataset on the Keystone consisting of
>     expired tokens etc.
>
>
>     Regards,
>     Neependra
>
>>     Cheers,
>>     Tristan.
>>
>>
>>     _______________________________________________
>>     Mailing list:http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>>     Post to     :openstack at lists.openstack.org  <mailto:openstack at lists.openstack.org>
>>     Unsubscribe :http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>
>
>     _______________________________________________
>     Mailing list:
>     http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>     Post to     : openstack at lists.openstack.org
>     <mailto:openstack at lists.openstack.org>
>     Unsubscribe :
>     http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20131217/3b8c6dd5/attachment.html>


More information about the Openstack mailing list