TEST 3 is now complete and the network is being brought down - big thank you to everyone who took part 
.
11 Likes
@Ross, Should we share the average number of Dropped messages here or will another topic be started about the test network being taken down and the request for the information?
1 Like
I think here is ok - keeps everything in one thread?
1 Like
I just have a couple of concerns with the how long this thread will get with people posting the information and the follow-up discussions. Plus having a new topic pinned right at the top of the forum saying test net 3 is over and what info who be a good way of informing the community. I just don’t want to see a people coming on the forum starting new topics of “Why can’t I connect” or “TEST 3 is down!”.
EDIT:I just noticed this topics title now says “Now Complete” at the end so that may be enough. 
3 Likes
Thanks, I think you have a lot of data again 
.
This test was also to see if the routing tables could be preserved even if all the chunks would be gone. This probably has to do with an RFC about Archive Nodes that’s coming 
. Is there any more info on that?
3 Likes
Dang it! I was just getting back to my testing and I can’t do much without a launcher. The network is still working for me just now - possibly reminents from the community but maybe not for long. I need Launcher logged in for my tests.
If testnet3 is disappearing any chance you can look at this problem building launcher for mock routing? Pretty please
1 Like
Ah sorry @happybeing - I missed that one - sure thing I will grab one of the guys and get back to you on that one.
1 Like
168 instances of dropped messages over the test period, ranging between 230 and 1 per instance.
Seems to be with different priority’s…I caught glimpse of priority >= 5 and priority >= 2
Not sure, I want to work out the average…probably should have a script to do it with that many.
2 Likes
Nothing greater than 3 here. Tip: fastest way for your brain to search node.log for this is doing a search for “priority >=” instead of typing the whole phrase out.
2 Likes
hmmm - okay - my bandwith sucked (latest after running too many nodes) and the log is taken from one/two nodes i was running on a laptop that might have had disconnects … don’t know …
If disconnections are a different category and not listed I’m not 100% sure if the “average” amount of lost messages is the best (/only important) criterion here Oo
Dropped 4 messages with priority >= 3.
354 priority >= 2
4 priority >= 3
80 priority >= 2
2 priority >= 5
3 priority >= 5
151 priority >= 2
1 priority >= 2
209 priority >= 2
236 priority >= 2
257 priority >= 2
1 priority >= 5
4 priority >= 5
1 priority >= 5
32 priority >= 2
8 priority >= 2
2 priority >= 4
19 priority >= 2
1 priority >= 3
15 priority >= 2
9 priority >= 4
5 priority >= 5
125 priority >= 2
399 priority >= 2
270 priority >= 2
33 priority >= 2
76 priority >= 2
33 priority >= 2
9 priority >= 4
197 priority >= 2
as you can see i had either very low numbers of lost messages or pretty high ones (which might also be disconnects …maybe …)
2 Likes
The network is still working for me just now - possibly reminents from the community
The magic of the SAFE Network resulted in a mesh style network that kept running even though devs took down the “ignition node”. 
Does this have an official name btw, does anyone know? Core devs fire up the network and nodes start joining, would we just call that Node_0 or something?
1 Like
I knew there is something unusual about this test taking too long 
5 Likes
These limits would have been fine, the problem is that 100MB and 500MB were used instead, which are too high for a network without coins and can explain the problems meet in test networks.