Magnuss

I had the same problem since some months, ok, not 41 minutes but still 2 minutes hanging on pooling. Now I have seen that I have in the table creature_respawn 63.000 records and in the gameobject_respawn table 18.000 records. After deleting them there is no waiting time. The core needs 18 seconds to start completely.

Thanks, it's working. As far as I understand ObjectMgr::InitializeQueriesData caches a lot of static data at core startup. I think the power of this cache can be only measured on very high populated servers. Regarding the fact that server memory is always precious I recommend to set default CacheDataQueries = 0.

Hello everybody Since about 2 or 3 months I have seen that the world server consumes about twice more RAM than before. Before it needed about 1.2GB at startup. Now, it's about 2.3GB. Maps are configured as unloading. Depending on player activities it's rising up to 3GB. I'm on an Windows 8.1 PC, tested with a fresh core and database: TrinityCore rev. 7b7dc1b50844 2018-05-12 13:59:34 +0200 (3.3.5 branch) Is that correct? And if yes, what was changes in the code that it needs about 1GB additional memory? What is the advantage?

@keonys Your screenshoots looks well so far. But in the morning a port check to this given IP told me they are closed. Don't know if your IP has already changed again or your server was down. May you want to send a screenshot of your router settings. But at the moment I can't say anymore. A possible way to check your network connection including port is using Telnet. If you have a Telnet client start it like that: telnet [IP] [Port] f.ex.: telnet 192.168.0.140 8085 to test a local connection or: telnet 176.164.12.14 8085 to test your public ip If your network is working the server will send some not human readable cryptic signs (about 30 signs) back to you. Press enter and in the TC world console appears a network error message like this: client [IP] sent malformed packet. Before you must configure your world.conf and enable Logger.network.

The host file you need only to change if you use a dynamic dns server like https://www.noip.com/ (I don't know if this service works for Trinity core.) First change on your server PC network setting so that you have a static IP and forward your ports to this IP afterwards. With a dynamic public IP you have to inform your friends about this IP every time it changes and they have to change the IP in the realmlist. (Or you use a ddns service). Additional you have to change the IP every time in the auth database realmlist and restart the server. In both config files (auth + world) there must be this (should be the default value): BindIP = "0.0.0.0"

127.0.0.1 for localAdress works well mostly. But is your public IP a static IP? Some provider in some countries change this on a daily schedule. Does your server PC has a static IP ? And did you forward both ports to your server ? (As SarcNBit asked?) Check your ports: https://www.yougetsignal.com/tools/open-ports/ (ATM they are closed for your given IP)

I know this part of the wiki. To tell the truth, exactly that's the point why we are using Windows. But 'we' means here 'some people' while some other people may prefers a more professional approach where stability goes over the wish to have always the latest gimmicks. (The term professional doesn't means in this case commercial.) Security patches, updates, testing, support, widely used release, hardware compatibility are some of the reasons to choose LTS. The funny fact is that Blizzard make it's own WotLK server software 10 years ago. And it was working. And I'm quite sure 2007 was no need for Ubuntu 17.x, wasn't it? Here are some nice words to LTS: https://askubuntu.com/questions/16366/whats-the-difference-between-a-long-term-support-release-and-a-normal-release?noredirect=1&lq=1 In addition to support, there are Development strategies that differentiate an LTS release: The base of the operating system, Debian, comes in three versions: Stable, Testing and Unstable. Normally, Ubuntu is based on Unstable; the LTS releases are based on Testing. Starting with 14.04 LTS, all new releases will be based on Debian Unstable. The Development effort for an LTS release in focused on providing a rock solid base, not only for customers who want the LTS release, but also for the next three Ubuntu versions to come. https://wiki.ubuntu.com/LTS Furthermore, we define the LTS to be: Enterprise Focused: We are targeting server and multiple desktop installations, where the average user is moderately risk averse. Compatible with New Hardware: We will make point releases throughout the development cycle to provide functional support for new server and desktop hardware. More Tested: We will shorten the development window and extend the Beta cycle to allow for more testing and bug fixing

ATM we are using Ubuntu 16.04 LTS for the mySQL Server (and a lot of other serving software) but still Windows for Trinity Core. As 16.04 is a LTS version it will stay for many years. I'm also interested why there would be a need to update a LTS version to some other non LTS version. IMHO for me it even doesn't make sense to update to next LTS 18.04 because it will be a brandnew version mostly less supported by companies, providers and community in the beginning.

Looks for me like a bug. In your case the variable ${_GIT_VERSION} in FindGit.cmake is empty. That why there is nothing written behind "Git version too old : " May you can wait till somebody else knows something or open a bug in the issue tracker with a link to this forum.

Before the error there should be a line like: Found git binary : C:/Program Files (x86)/Git/cmd/git.exe You can test it in a windows console with : "C:/Program Files (x86)/Git/cmd/git.exe" --version That's more or less what the CMake script is checking to determine the version.

CMAKE fails in FindGit.cmake: # make sure we're using minimum the required version of git, so the "dirty-testing" will work properly ensure_version( "${_REQUIRED_GIT_VERSION}" "${_GIT_VERSION}" _GIT_VERSION_OK) # throw an error if we don't have a recent enough version of git... if(NOT _GIT_VERSION_OK) message(STATUS "Git version too old : ${_GIT_VERSION}") message(FATAL_ERROR " Git was found but is OUTDATED - did you forget to install a recent version, or setting the path to it? Observe that for revision hash/date to work you need at least version ${_REQUIRED_GIT_VERSION}") endif() endif() CMAKE can't find a value for ${_GIT_VERSION}. Seems to by empty in your case. Don't know why. May it helps to reinstall Git Extensions and/or CMake following exactly the guide in TC Wiki. BTW: Can you show the full message log of CMake before the error message ?

No

The error found was: Git was found but is OUTDATED So, what is your GIT version ? You can open a Git Bash console in Git Extensions and type there git version.

No official support by what ? CMake, Boost, TC ? VS is updating itself quite fast. Today it's already 15.3.2.

I have a lot of warnings "Unknown compiler version - please run the configure tests and report the results" in VS log. I'm not sure but I think it's since the last VS update and wasn't before. However it compiles without errors. (CMake is also without errors) After a Google search I think it something with Boost libs. But I already use the latest stable 1.64 version. The VS is version 15.3.1. Somebody else also have this warnings ? Are they new ?

Trinity core is an evolving work for about 10 years. Therefore a lot of things has changed in the past (even the wiki software) and so it's not possible to hold the wiki every day up to date. If you have questions please use the forum or if you think some parts in the wiki needs to be better descripted you can do this even yourself if you want. (That's the idea of a wiki btw.) I have changed today the page for GM commands. Hope it will help you better. https://trinitycore.atlassian.net/wiki/spaces/tc/pages/2130065/GM+Commands And I agree that may some parts are confusing like the fact that the database part follows after the core starting part. Maybe somebody else will have a look on that. But in general I can say the wiki contains really a lot of important and detailed information.

Thanks for the answers. Now it's more clear for me.

My knowledge of c++ is limited and I would like to ask how that works: The declaration of creating a gameobject is in GameObject.h (line 687) like that: bool Create(ObjectGuid::LowType guidlow, uint32 name_id, Map* map, uint32 phaseMask, Position const& pos, G3D::Quat const& rotation, uint32 animprogress, GOState go_state, uint32 artKit = 0); After phasemask the function is asking for a struct Position. But in the script for spawning gameobjects (cs_gobject.cpp, ) there is a piece of code like that: Player* player = handler->GetSession()->GetPlayer(); Map* map = player->GetMap(); GameObject* object = new GameObject(); ObjectGuid::LowType guidLow = map->GenerateLowGuid<HighGuid::GameObject>(); G3D::Quat rot = G3D::Matrix3::fromEulerAnglesZYX(player->GetOrientation(), 0.f, 0.f); if (!object->Create(guidLow, objectInfo->entry, map, player->GetPhaseMaskForSpawn(), *player, rot, 255, GO_STATE_READY)) { delete object; return false; } As you see instead to send a position struct there is a pointer to player. How does this trick work ? There is no compiler complaint and I also couldn't find an overload.

Quests can be done just only by editing the database. No recompile nor a client patch is necessary. Basic quests attached to a NPC works mainly like that: 1. Use or create a NPC and set its npcflag to 3 (Gossip and Questgiver) 2. Create a quest template. 3. Make a link in creature_queststarter and creature_questender

No, you don't need a client patch for that. But sometimes you need to delete your WoW wtf cache folder. All texts and some other parameters are cached. However, if the text is missing at all you may have done something wrong.

Perfect. Thanks. If it would be on me I also would add the information that References points to reference_loot_template only. Because that's new afaik.

If references points to reference_loot_template ONLY than the SQL example above can't be declared as an example of a self-reference. Inserting 21215 in field reference in creature_loot_template than would (as I understand you) point to reference_loot_template. Where is the self-reference ? Selfreferencing can be IMHO done only inside reference_loot_template. In the other loot_template tables the worst thing you can do is making a loop. So I think the wiki text isn't correct and describe the situation last year before changing a lot of stuff in loot tables.

But in this case the SQL in the wiki doesn't make sense for me: INSERT INTO `creature_loot_template` (`Entry`,`Item`,`Reference`) VALUES ('21215','0','21215'); That can't be a self-reference.

In the wiki the field Reference is descripted as follows: Template reference asks core to process another loot template and to include all items dropped for that template into current loot. But another loot template from where ? From the same table or from reference_loot_template table ? To which one is it linked ?

hehe :-)