12:20 < bridge> [teeworlds] <fokkonaut> Is there anything new about the master server problem in 0.7?
12:20 < bridge> [teeworlds] <fokkonaut> That servers can lose connection to it
12:20 < bridge> [teeworlds] <fokkonaut> (for me: when i login to rcon, because many rcon commands are sent then)
12:38 < bridge> [teeworlds] <fokkonaut> okay so, i set MAX_RCONCMD_SEND=4, it was 16 before
12:38 < bridge> [teeworlds] <fokkonaut> now it seems to work :D ?!?!?!??!!
12:39 < bridge> [teeworlds] <fokkonaut> because less traffic at once, the server doesnt lose the master
12:41 < bridge> [teeworlds] <fokkonaut> @Dune @oy i think you guys should look into this. Mods can and will always have more rcon commands than vanilla. For me this issue came over time as I added more and more rcon commands. If you have (guessing now) 80+ rcon commands or 100+ as my mod/ddnet it obviously causes a lot of traffic when logging into rcon
13:00 < bridge> [teeworlds] <Dune> Is there an issue?
13:42 < bridge> [teeworlds] <Learath2> Rcon flood has always been an issue
13:43 < bridge> [teeworlds] <Learath2> But its usually an issue with the clients getting dropped because of a full buffer
13:43 < bridge> [teeworlds] <Dune> I mean, is there a github issue
13:43 < bridge> [teeworlds] <fokkonaut> no
13:43 < bridge> [teeworlds] <fokkonaut> I will create one
13:43 < bridge> [teeworlds] <Learath2> I didnt think it would cause a register issue
13:43 < bridge> [teeworlds] <Learath2> Ohh it could be flooding the sendqueue
13:43 < bridge> [teeworlds] <Dune> a well-detailed issue that corners the problem is a good start to invite a generous contribution
13:43 < bridge> [teeworlds] <fokkonaut> @Learath2 same can happen on ddnet7 with votes, on map changes/difficulty change
13:44 < bridge> [teeworlds] <fokkonaut> but not on ddnet6 or any ohtter 0.6 mod
13:44 < bridge> [teeworlds] <Learath2> Register should honestly not be using the same netserver
13:44 < bridge> [teeworlds] <fokkonaut> maybe that would fix it already
13:44 < bridge> [teeworlds] <fokkonaut> but the weird thing is, that it cant even find the masters back
13:44 < bridge> [teeworlds] <fokkonaut> even when all commands are sent, it just cant find the master back again
13:45 < bridge> [teeworlds] <fokkonaut> just a server restart helps
13:45 < bridge> [teeworlds] <Learath2> Check a wireshark, see whats going on
13:45 < bridge> [teeworlds] <fokkonaut> i cant, its on chillerdragons vps
13:46 < bridge> [teeworlds] <Learath2> It might be that during the flood a token message goes missing and the server keeps dropping the new messages because wrong token
13:46 < bridge> [teeworlds] <fokkonaut> that has to be it
13:46 < bridge> [teeworlds] <fokkonaut> that explains why its not in 0.6
13:46 < bridge> [teeworlds] <fokkonaut> there were no tokens back then
14:53 < Oy> @fokkonaut could you add the register messages of the server in the issue?
14:54 < bridge> [teeworlds] <fokkonaut> sure
14:56 < bridge> [teeworlds] <fokkonaut> @oy, added it
15:09 < Oy> ok
15:10 < Oy> shouldn't be the token though, as every message does a new token request
15:10 < Oy> maybe same problem the version server is having
16:32 < bridge> [teeworlds] <fokkonaut> I am not familiar with the other Problem
17:11 < bridge> [teeworlds] <redix> Did this happen on servers other than chillers vps?
17:15 < bridge> [teeworlds] <fokkonaut> Maybe chiller vps is not that fast in networking?
17:15 < bridge> [teeworlds] <fokkonaut> he also runs other things there like a minecraft server
17:16 < bridge> [teeworlds] <heinrich5991> it happened elsewhere, too, I got a package dump that showed an unrelated problem
17:16 < bridge> [teeworlds] <heinrich5991> a tcpdump would be really valuable
17:17 < bridge> [teeworlds] <redix> Well that server tends to be a bit laggy and has packet loss sometimes (checked from several connections) but it still should not happen that frequently
17:17 < bridge> [teeworlds] <redix> Okay
19:35 < bridge> [teeworlds] <Dune> Depending from where I am that server has high packet losses yeah