Thyth · Apotheosis Incarnate

  • Thyth
  • Member since
  • Last active (8 visits)

Comments

  • Aiming for account and browser systems online on Tuesday or Wednesday night.
  • One of the two hardware issues I was seeing has been resolved. Still running a few tests on the second issue (doesn't seem operationally relevant, but making sure). The very last hardware bits are arriving tomorrow. I'll be installing the hypervisor and administration OS this weekend, and finishing a set of security…
  • Looking like the week of the 14th is most realistic for getting this box online. I've been "unscrewed", but a couple of hardware issues came up during testing that I need to understand or resolve. Neither look show stopping, but I still need to understand impact. Given the long delays, I'm going to attempt standing up a…
  • Nothing is ever quite as smooth as I'd like. At last, today, almost all of the new server hardware has arrived (inc. server itself, CPUs, RAM, drives), and has been tested and verified as functional. I was going to begin hypervisor and OS installation this evening, after which restoration of the retired server backup could…
  • Downtime is going to be at least an extra week. Aiming to bring services back online the week of August 31st.
  • I'm beginning the downtime window as promised. See you on the other side in approximately one week.
  • Potentially silly solution: create your different gametype map versions with a slightly different name but otherwise identical non-sphere map contents.
  • What graphics processor is your computer running, and do you have the latest drivers?
  • The TribesNext account and browser backend systems are now back online.
  • Turns out the server really is down. Due to some over-zealous security settings, an LDAP software configuration issue, and an extended power outage, the earliest I'll be able to bring TribesNext account/browser services back online will be on the evening of this Saturday the 20th (when I can get physical access).
  • Good to hear. I've added a note in the FAQ answer regarding these error messages clarifying that LAN servers must use RFC1918 addressing schemes.
  • What is your LAN IP addressing scheme? 192.48.26.0/24 network address in your LAN connection error corresponds to an Internet routable segment of Boeing's network. You cannot use Internet routable non-RFC1918 IP addresses in combination with NAT, or it will appear as if a man-in-the-middle IP routing attack is taking…
  • FYI, it's possible to extract a T2 map from a T2 .rec file of someone playing on it -- assuming you're able to play back the .rec demo (if the map has custom shapes for buildings or terrain, you need those for playback/reconstruction).
  • I've copied code from the TribesNext SVN repository to Github: https://github.com/TribesNext/t2-scripts This includes the incomplete in-game GUIs, but is intended for developers, not general users at this time. Heat has some bandwidth to work on this stuff. I'll be accepting pull requests, and if we end up finishing the…
  • You're probably better off doing this by GUID than IP. GUIDs are unique numeric identifiers part of each account, thus you could whitelist players, rather than IP addresses. I had to re-implement bans when building the TribesNext version of the game server connection process, so the code for processing GUID/IP bans is in…
  • That kind of provision in an EULA is legally unenforceable for reasons I discussed already. See: It's the same kind of legal smoke and mirrors the NFL tries to play with trademark law around "the big game". It's a house of cards, but nobody has actually gone through the motions to blow it down.
  • I think these issues are related, but I don't think it's a packet ordering issue. There's no hard deadline (other than the 30 second timeout) for the arrival for a move related message packet, so while it might average out to 1 move per simulation step, at a local time interval, there is the possibility for staggering…
  • After testing the second patch, and re-reading the Torque code, I realized that the issue was due to uneven bucketing of incoming client movement data into simulation steps. My next idea was to add two fields to each client to track how much further ahead, or back in time they are relative to true server time. If a client…
  • Defense Turret was beaten because there are two constant "escape hatch" codes that it uses if you failed paired consensus more than 5 times, or were otherwise not paired with another player (e.g. odd number of players connected). It's probably derived from hashing some part of the T2 binary (since there were different…
  • Those two quotes are at odds. You either own the rights to your work and can choose how it is used by others, or it doesn't belong to you and you get no say. I take no offense, I'm just unequivocal that your first statement (that I originally quoted) is utterly wrong, and that I vehemently disagree with the philosophical…
  • Even free and open source software requires respecting the property and creative rights of its authors. Developing free (as in libre) software should always, and will always be a choice. TribesNext is not free software, but I'm sure you already knew that when you read and accepted our end user license agreement. We wrote…
  • This is so wrong on so many levels. You always own rights to your creative works, which are retained until 50 to 70 years after the creator's death. This applies to literature, art, software, et cetra. Creators have a right to control the terms under which others can benefit from that creative work, be those terms…
  • Do not install the tournamentNetClient.vl2 on dedicated servers.
  • FYI, scheduling latency isn't the cause of issues with T2 on symmetric-multi-processing systems. Rather, it's the use of the CPU internal time stamp counter (RDTSC instruction). Some multi-processing systems don't make any attempt to keep time stamps synchronized between execution units (especially because it's just…
  • One thing you may try (especially if the server is running on a recent Intel or AMD multi-core CPU) is lock the process to one core (or switch from RDTSC to HPET timing mode in the game). If the timestamps differ between execution units on the server, the simulation will jitter as the process is scheduled across different…
  • SCP runs custom map packs. See this thread on their forum for details: http://www.t2forums.com/index.php?topic=166.0
  • An updated version of checkver.cs is now available from the link above. Heat Killer had done some tests and helped track down a couple issues. A version of this script is now live on Snap Crackle Pub, and will eliminate people still running HM2 by reverting to the older patch.
  • I put together a script to check for clients running a new enough version of TribesNext with the mini-browser client created for the tournaments: http://thyth.com/tn/checkver.cs It's totally untested, but, if you install that script on a T2 server (RC2a+ on server side), it will prevent anyone without RC2 &…
  • Oh, here we go again. Another battle of wit that I choose not to witness. Thread locked.
  • Firstly, I'm unconvinced that the TribesNext boards are the right place to discuss grievances with particular servers, but as long as people behave themselves, I don't care about topics of discussion. In order to keep the forum experience consistent for everyone, I will continue to lock threads that become a hostile battle…