Test Server Feedback

Hey haba, thanks for the update.

I got an unrelated question though: In fanfest we saw the presentation where...veritas I am thinking but probably wrong...anyway...someone showed that thin clients are a way to test with large amounts of ships, instead of masstests.

Since those clients however are far more predictable than human behavior and don't act in the same way, I was wondering:

Will this have an impact on the frequency of masstests? There hasn't been a masstest in a while anyway :(

~Gotta get dem SPs~

CCP Veritas was speaking at fanfest about our new NoClients, which are similar to thinclients in its field of application but quite different in its setup. Thinclients are stripped down clients without any graphics, while NoClients are simulated in the proxy layer of our server - without any clients at all. This means that NoClients are faster and more efficient (no extra hardware needed), but on the other hand we have to be more careful with the interpretation of results and we are not sure yet about some of the boundaries.

I do not expect a lot of changes for mass tests in the near future. They are for sure still needed. Thinclients and NoClients are good in testing a specific feature, but mass tests are much better to get a broad coverage - both for performance and stability. Mass tests are also used in some cases to validate results from thinclients.
The number of mass tests was indeed a bit lower in the last time, which was mostly caused by lack of features, which needed to be tested in a mass test. There will be for sure mass tests for the Kronos release, but I'm not sure yet how many. Later in the year we will very likely need quite a few more mass tests, as soon as the Dogma rework and Brain in a Box will be ready for testing.
About combining thinclients and mass tests: We thought about this before, but so far we have not found a good use-case for this combination. I'll keep my eyes open :)

FML! 2nd time internet crashed as I pressed post and it deleted what I typed and draft is of a totally different post...WTF



Thanks for the response and the expanded explanation!

I can especially understand how masstests are important for both server performance and client performance tests, with all of us running different setups.
And the features that could be tested with so many was lacking I admit.

But since Kronos is so industry focused, what could need masstesting? Or are there some features coming we don't know about yet?

Will dogma/brain in a box have any impact at all client side, or is it only server side improvements?

500 abaddons instead of 50 officers could be fun!
Now get back to mirroring!

@Zappity,
Masstests are still needed, as CCP Habakuk just said.

When 100m USP happens, everyone will fly the same things and it will be even worse than DeadSpace mirrors, fun for a little while and it just becomes ridiculous after that.

Very Happy to hear that mass test wont be going away

As per to 100M SP i think that big amount is ridiculous but i will like say an increase in awarded SP will be nice, as often the purpose of test server is to test out features. Let me just say thnx to test server i have saved months over months of useless training as I have first tried out something new here see how i like or not am i effective or not and than go train that on TQ.

But the timing of masstest sometimes make it impossible to login, I understand we reward who come test stuff for us but pretty please shower some extra love on ppl who do play on sisi regularly and test different stuff. :)

The mirror can take up to three days to snapshot and deploy I believe. They have to cleanup the mirror a bit, adjust moon materials content, and removing POSes soo I would imagine either tomorrow or Thursday if I had to guess.