Performance tests

[nbusch]

Hey there,

I am working on my engineer thesis. My task is to make adaptive and performance test of about 4 clouds using OpenKM. I downloaded the community edition and adapted it on Amazon EC2 instance - Ubuntu 12.04 64 bit. Everything adapted well. Now I need to make performance tests. Does anyone have an idea for performance tests? I ask you because I was thinking about them since yesterday and nothing clever come to my mind. My first idea was to upload different sizes of files and different ammount of files and check the upload time, but the problem is I have no idea how to check the upload time. Could you please help me in my performance tests? How could I make them?

Have a nice day and thank you very, very much in advance!

[jllort]

First must take in mind what are you trying, document steps are:
1- Upload ( bandwidth problem )
2- After upload document goes to batch indexer queu where will be procesed internally by text extractors that use a lot of cpu resources
3- When you got a lot of documents ( search speed can get different perfomance )

You can check performance from several ways:
1- UI ( upload stress ), some tools can help you on it ( http://www.neotys.com/product/overview-neoload.html ->aplication neo load), the google query is ( gwt performance test )
2- With indexer stopped ( at crontab you have a task called Text Extractor Worker, simpy disable it ) changing cron time can make it more agressive, to use all hardware take a look at http://wiki.openkm.com/index.php/Applic ... extraction -> the idea is use all cpu's ). Test of hardware load. Take in mind if you index pdf needs a lot of cpu to parse document ( with ms office less ) and text files are instantly processed
Example of configuration for two cpu where batch is the number of files procesed each time crontab task is started

managed.text.extraction=true
managed.text.extraction.batch=80
managed.text.extraction.concurrent=true
managed.text.extraction.pool.size=2
managed.text.extraction.pool.threads=2
managed.text.extraction.pool.timeout=1

3- You can use scripting to create files dinamically http://wiki.openkm.com/index.php/Scripting_-_OpenKM_6.2 you should get minimal skills about openkm api for it to create folders and documents into take a look here http://doxygen.openkm.com/ and specially OKMFolder (http://doxygen.openkm.com/openkm/d0/dd8 ... older.html) and OKMDocument (http://doxygen.openkm.com/openkm/de/de7 ... ument.html)

You can create in few hours some milions of docs ( the idea is create text files with different contents ) -> with it you can get some idea about hardare comparision between cloud providers.