Lt. Commander
Join Date: Dec 2007
Posts: 120
# 11
11-23-2011, 01:26 PM
Quote:
Originally Posted by Enterprise-D View Post
20 minute test??

How did you get any reliable data from that short test?
As a programmer, if I was testing this, I wouldn't need actual users to test it out... I'd just create a bunch of simulated logins (actual logins as far as the server is concerned) during that time, and analyze the results on that. As users would get affected during that time, just inform people that the test is going on, do the test, some might get stuck during that time, but all I care about in monitoring the full queue.

20 minutes is plenty of time to perform a test, if done right
Lt. Commander
Join Date: Dec 2007
Posts: 120
# 12
11-23-2011, 01:41 PM
Quote:
Originally Posted by Rehpic
We are going to turn on the login queue briefly this morning on Tribble to test it out. If you happen to get in the queue please stay in rather than logging out and trying again later. This will help ensure that if we need the login queue during f2p launch that it will run smoothly.

I will post again when the testing is complete.
Its not the question IF you need it it is for how long you will need it.
Lt. Commander
Join Date: Dec 2007
Posts: 120
# 13
11-23-2011, 02:56 PM
Quote:
Originally Posted by Sardoc
As a programmer, if I was testing this, I wouldn't need actual users to test it out... I'd just create a bunch of simulated logins (actual logins as far as the server is concerned) during that time, and analyze the results on that. As users would get affected during that time, just inform people that the test is going on, do the test, some might get stuck during that time, but all I care about in monitoring the full queue.

20 minutes is plenty of time to perform a test, if done right
The server might handle the queue just fine, but the client also needs to work in the queue scenario.
One issue they had in DCU was that in some cases the client wouldn't go into queue mode and make the client freeze on loading screen or that the server took too long to give an update regarding the queue so that the client terminated the connection due to time out and thought it had been disconnected.

Hammering the server with simulated logins is one way of testing it, but nothing beats the same amount of logins from real players that gets frustrated at ending up at # 1000 or think that they are stuck so they log out and back in again several times to get to a lower # that will generate allot more stress on the login queue than simulated ones.
Lt. Commander
Join Date: Dec 2007
Posts: 120
# 14
11-24-2011, 05:47 PM
Think of it this way: You have an old car that has been sitting for 2 years, you put some fuel in it and you go to fire it up. The test was most likely to test if the system would even start up. You wouldn't leave it running for very long because it could break something, but you know it works.
Lt. Commander
Join Date: Dec 2007
Posts: 120
# 15
11-27-2011, 08:27 AM
Quote:
Originally Posted by Enterprise-D View Post
The server might handle the queue just fine, but the client also needs to work in the queue scenario.
One issue they had in DCU was that in some cases the client wouldn't go into queue mode and make the client freeze on loading screen or that the server took too long to give an update regarding the queue so that the client terminated the connection due to time out and thought it had been disconnected.

Hammering the server with simulated logins is one way of testing it, but nothing beats the same amount of logins from real players that gets frustrated at ending up at # 1000 or think that they are stuck so they log out and back in again several times to get to a lower # that will generate allot more stress on the login queue than simulated ones.
Simulated logins fill up the queue, at which point you can use real logins that you own to test out how the client handles it. So while your real clients are in the queue, you kill the simulated logins (ie the player hit 'x' to quit the game, leaving the queue). Queue decreases, you get in the game, and full test is complete. Relogging multiple times randomly with simulated logins + some more test clients can be tested as well. Still doesn't take longer than 20 minutes.

I did system testing for a couple years, if done properly, it's very easy to simulate real conditions, and generally that's what you always should do before deploying any software/system.
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


All times are GMT -7. The time now is 07:29 AM.