Servers workarounds

[Nelsona]

Probably this week-end some servers still left with XCv23 and plain original net-driver will go switched to XCv24b and XC_Ipdrv.
For XC Players downloads can be boosted directly from game-server. If certain things are going badly messed up causing some map to be unusable, let me know.

Servers were operational pretty well so far but... certain events regarding to networking part last time were causing some annoying behavior of servers and then it might be time to do something here. I'm expecting some list with MonsterHunt "maps" to be removed as long as some of them are just full crap and nothing more. It's not an offense pointed at mappers, it's just a reality observed by players all the time. Some of these were done in 2005-2009+ years by n00b kids and they should be taken in account as risk factors for stability, unless you like to reboot/crash server all the time - I'm not volunteer for being a server guardian.

Updating will need a bit of outage.

When 469c will be out, then it won't be my problem any testing stuff, as long as it won't run on my machines as it was announced - unless you have an Android version of UT .

[BluEyedBitch_{HoF}]

Wait... I hope I won't have to install that hideous driver again ! It was so very wrong for my computer ... Tell me I can keep my v436 !! Or even that other one works for me , uh... v451 or somethin...

[Nelsona]

Player using 436 mainly is able to play game - What if source-code would not be released ? Updating is not mandatory unless it's solid and stable.

I'm currently using 436 and XCGE v24b for client. I was playing generally out of problems, unless some admins don't really know what they do and their servers are empty due to various factors keeping players away (screwed stock, other funky natives, messed up redirect - various mismatch things, borked maps, etc.). For Hi-Res textures I don't recommend those very HD ones 2k etc, they are consuming memory which is not that welcomed in UE1.

MonsterHunt server with Invitation has been upgraded and tested a few time. To me it looks operational. It uses now a conformed MH by Buggie, but I don't see too many errors solved, I'm not sure if it worth keeping said updated MH package only for very few fixes.

[Nelsona]

In discussion NWIII Invitation server
All right... Until the list with removals will get ready I dropped away those maps which are not having an End - junks more exactly. I don't know how are added maps in server but if these cannot be ended I don't see their purpose. The rest of Stock maps are "OFF-Topic" for a MonsterHunt server and then I dropped them away too.

And now probably a few other maps with bad settings or BSP problems can be removed too but... it depends on players desire...

[Nelsona]

I see that server looks operational since nobody pointed anything else to be adjusted/changed - good to know.

Recommend me next server for adjusting settings.

[SC]-[WARTZ_{HoF}]

Evil Sniper.

[Nelsona]

There isn't Evil the dude delegated to add/remove maps ? Or we talk about anything else ?
Or maybe some lords of UE1 updates can do something else ? That server months ago with the same settings has worked like a charm. Now it logs missing skins and has sudden connections issues which weren't there before any 469 thing. Do we need assistance from NfoServers if it's about some changes at Networking chapter affecting VDS ?

[SC]-[WARTZ_{HoF}]

Could be NFO that has issues. Maybe we need to move the VDS to another location.

[Nelsona]

To me this is not smelling well. Let's see something older.
Some of U227 bugs were left in game because "they could not be recreated" during working stage so they stay there splitting users in parts: U227 users and U227 more or less haters.
You see ? I think in 469 situation will be the same. Not everyone can report bugs at GitHub because of Browser restrictions and then... game is old and not mandatory for running only on newer machines as some boys decided or are thinking. I think this will have a less nice end, 469 players are going to be firewall-ed as long as they would use 436 if 469 would not exist.
After working to do a stable server, a new series of issues are giving now headaches. What for after all ? Because of new ideas BAD executed ? Because of structures changed for no reason ? Because of hidden implementations which are not an answer at anything ?

Yeah, let's clap hands because something toward states changing was fixed. Amazingly my pawns are/were not having such issues but servers are facing new problems. Either way not every single map is running in 469b and we do have damage coming this way. I don't think these are my fault, I did not write anything for 469 except recommending some adds for Editor which I see ignored. If I'm ignored I'll ignore this 469 too as long as I don't have issues with 436 playing in a 451 server.

When they decided to reconstruct navigation chain with empty nodes and void nodes I'm curious what anything else was done this way and causing only pain. Excuse me but these are not my ideas or my contributions, to me this is not what a patch is supposed to do and this way is not a COMPILER problem, is a CODER problem. ReachSpecs and navigation chain must be well constructed depending on map state and it's not a must do this everywhere in the same way. PlayerStart from void remapped would cause only damage and without helping with anything - I worked to patch maps at this point. I have never expected such evil move in a patch, it's insanity. These gaps should be EXCEPTED and NOT remapped for spawning player into void.

The only "attempts" for solving problems would be switching to XC_IpDrv and XCGE 24. As long as a 469c is expected, logically 469b has issues or else a 469c would not be under construction, demonstrating that 469b is not really ready yet.

Back at servers:
I reverted back TEXTURES files from Blue's server which is version 451, because someone added those from 469 and I've read something funky in logs. I recommend leaving all alone or replacing all files if needed and not only fragments of them.

Evil Server - Perhaps it's time to see what "Evil" term means. There can be installed utlogin2 whatever and collected all data about 469 players and then... firewall-ing them all. Then, let's see if server still drops connections...

Update note
That "login" stuff isn't logging anything readable... I'm going to adjust a few settings such as adding back all "window" and "render" stuff which is not really visible in a dedicated server but maybe has something to do with new "networking" or it has nothing to do as long as EUTSource doesn't use these and it's capable even to record games server-side.

[Nelsona]

Okay, assets configured in old style (declaring render canvas and all that invisible-for-server objects) doesn't seems helpful. As long as server is bombed with those "alien" connections, it's not stable. Connections are reset and is visible in NSCLogs that in a moment all player has left server, returning after a few time.
These problems are over my capabilities, I don't get what's happening in networking as I don't see data in a human readable format.
Updating to 469b ?
Perhaps if a 469c it's a work in progress, this means that 469b (which is out and it does damage) it's not finished.
Until a boy to show up with a solid network driver or a good patch you can live with these as they are, shortly: Good luck with the server and the patching process.

All was good a few weeks ago - now it's crap, and I did not do anything before these connections issues which don't seem related with network timers configured.

Here are two options:
- trying to figure if exist a solution through firewall - I need to study these well;
- finding a coder capable to write an IpDrv driver with a configuration delegated to discard a connection which has some specific evil string data - ignoring certain version or a connection including matching data from a TXT file;
- anything else ?