[Deploying Sakai] [Building Sakai] Load Testing CLE

Warwick Chapman warwickchapman at gmail.com
Fri Nov 1 14:56:58 PDT 2013


Thanks Matthew.  I have a 'class' of 250 logging in for the sole purpose of
completing two MCQ's (30 mins) and one essay question 1 hour.

Speaking of which, last weekend we ran this, also for about 250 people, we
had a lot of issues with people submitting their essays after 1 hour of
typing them... any suggestions how to improve that?

-- Warwick Bruce Chapman | +27 83 7797 094 | http://warwickchapman.com


On Fri, Nov 1, 2013 at 11:51 PM, Matthew Jones <matthew at longsight.com>wrote:

> No, those aren't the same tests. You aren't going to find any tests that I
> know of that are going to provide you with any accurate real world
> simulations of a Sakai instance. There are some load tests out there that
> perform simple tests with jmeter/grinder that Chris mentioned to see how
> much load the portal or a few tools can take as far as read on their first
> page is concerned with X number of users, perhaps this is enough, but will
> still take some effort to stand-up.
>
> I tried writing some things on neustar/browsermob a few months ago and
> wasn't very successful, mostly because of all the frames Sakai has and some
> security protection implemented on form posting that you'd have to account
> for. It didn't make recorded scripts very useful. They all had to be hand
> modified/written and it was producing unreliable results (sometimes running
> and sometimes failing on the same script). Starting from the /pda portal
> would have been a good idea. Maybe whatever bug I was having is better now,
> but it was frustrating for some reason.
>
> Most of the real load related issues you'll run into are the hard to
> predict edge cases rather than typical everyday usage. The class with 100
> students and everyone trying to take an exam with some specific set of
> publish options and questions at the same time or everyone logging in to
> view their grades.
>
>
>
>
> On Fri, Nov 1, 2013 at 5:08 PM, Chris Kretler <ckretler at umich.edu> wrote:
>
>> Hi Warwick,
>>
>> There have been a number of efforts to openly load test sakai.  Some of
>> them used The Grinder:
>>
>>
>> https://confluence.sakaiproject.org/display/PERF/Community+Performance+Testing+Resources
>>
>> https://source.sakaiproject.org/contrib/sakaigrinder/trunk/
>>
>> Others have used jmeter:
>>
>> https://confluence.sakaiproject.org/display/QA/Custom+Test+in+10+Minutes
>>
>> I haven't touched the grinder scripts since 2009, so they would likely
>> require updating before running.
>>
>> You could also look into leveraging browsermob to execute Selenium
>> scripts at large scale:
>>
>>
>> http://www.neustar.biz/enterprise/web-performance/what-is-load-testing#.UnQUxBBGYfQ
>>
>> This does cost money, but the selenium scripts are probably the easiest
>> to create, assuming they need to be created from scratch.
>>
>> Regards,
>> Chris
>>
>>
>>
>>
>> On Fri, Nov 1, 2013 at 3:15 PM, Warwick Chapman <warwickchapman at gmail.com
>> > wrote:
>>
>>> Howdy
>>>
>>> I've searched and only found information relating to load testing Sakai
>>> 3.  Are there some scripts for load testing Sakai 2?
>>>
>>> I want to run some heavy load tests tomorrow to help me tune my
>>> AppServers for D-Day on Sunday - is this possible?
>>>
>>> Thanks very much.
>>>
>>> -- Warwick Bruce Chapman | +27 83 7797 094 | http://warwickchapman.com
>>>
>>> _______________________________________________
>>> production mailing list
>>> production at collab.sakaiproject.org
>>> http://collab.sakaiproject.org/mailman/listinfo/production
>>>
>>> TO UNSUBSCRIBE: send email to
>>> production-unsubscribe at collab.sakaiproject.org with a subject of
>>> "unsubscribe"
>>>
>>
>>
>> _______________________________________________
>> sakai-dev mailing list
>> sakai-dev at collab.sakaiproject.org
>> http://collab.sakaiproject.org/mailman/listinfo/sakai-dev
>>
>> TO UNSUBSCRIBE: send email to
>> sakai-dev-unsubscribe at collab.sakaiproject.org with a subject of
>> "unsubscribe"
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://collab.sakaiproject.org/pipermail/production/attachments/20131101/1e49c42f/attachment.html 


More information about the production mailing list