[HELP] Adding a Second Cluster (not second channel)

NickHough · May 24, 2012

Hey,

I am trying to add a second cluster to try expand my server, i have it all working (2 dedicated servers completely fit to work for 2 clusters)

The only problem is when you get to character selection you can't actually enter the world

I origionally followed this guide http://forum.ragezone.com/f651/add-new-cluster-channel-imansos-704381/
and it didn't work, well not completely, i have the same start up as him with a slight twist

My start up is as followed:

Main Cluster:

1-AccountServer.exe
2-Certifier.exe
3-DatabaseServer.exe
4-CoreServer.exe
5-LoginServer.exe
6-CacheServer.exe
7-WorldServer1.exe
8-WorldServer2.exe
9-WorldServer3.exe

Secondary Cluster:

1-DatabaseServer.exe
2-CoreServer.exe
3-LoginServer.exe
4-CacheServer.exe
5-WorldServer1.exe
6-WorldServer2.exe
7-WorldServer3.exe

My ini's are as followed on my Cluster 2:

DatabaseServer.ini

Code:

Sys		2
ResourceFolder	"C:\Program Files\Flyff V17 working\Resource"
Account		"Main Server IP"
DSN_NAME_CHARACTER    "character01"
DB_ADMIN_ID_CHARACTER    ""
DSN_NAME_LOG    "log01"
DB_ADMIN_ID_LOG		""
DB_PWD_LOG	""
DB_PWD_CHARACTER	""
DB_PWD_BACKEND	""
DB_PWD_ITEMUPDATE	""

CoreServer.ini

Code:

Sys	2
Database	"127.0.0.1"

201
{
	1	0	0	30	30	00	00
	2	0	0	5	5	00	00
	3	0	0	5	5	00	00
	4	0	0	5	5	00	00
	21	0	0	1	1	00	00
	22	0	0	1	1	00	00
	23	0	0	1	1	00	00
	24	0	0	1	1	00	00
	25	0	0	1	1	00	00
	100	0	0	1	1	00	00
	101	0	0	1	1	00	00
	102	0	0	1	1	00	00
	103	0	0	1	1	00	00
	121	0	0	1	1	00	00
	122	0	0	1	1	00	00
	123	0	0	1	1	00	00
	124	0	0	1	1	00	00
	125	0	0	1	1	00	00
	126	0	0	5	5	00	00
	127	0	0	5	5	00	00
	128	0	0	5	5	00	00
	129	0	0	5	5	00	00
	130	0	0	12	5	00	00
	131	0	0	10	10	00	00
	132	0	0	5	5	00	00
	133	0	0	5	5	00	00
	134	0	0	10	10	00	00
	135	0	0	5	5	00	00
	151	0	0	1	1	00	00
	152	0	0	1	1	00	00
	153	0	0	1	1	00	00
	154	0	0	1	1	00	00
	155	0	0	1	1	00	00
	156	0	0	1	1	00	00
	157	0	0	1	1	00	00
	158	0	0	1	1	00	00
	159	0	0	1	1	00	00
	160	0	0	1	1	00	00
	161	0	0	1	1	00	00
	162	0	0	1	1	00	00
	163	0	0	1	1	00	00
	164	0	0	1	1	00	00
	165	0	0	1	1	00	00
	166	0	0	1	1	00	00
	167	0	0	1	1	00	00
	120	0	0	1	1	00	00
	200	0	0	1	1	00	00
	201	0	0	1	1	00	00
	202	0	0	1	1	00	00
	203	0	0	1	1	00	00
	204	0	0	1	1	00	00
	205	0	0	1	1	00	00
	206	0	0	1	1	00	00
	207	0	0	1	1	00	00
	208	0	0	1	1	00	00
	209	0	0	1	1	00	00
	211	0	0	1	1	00	00
	212	0	0	1	1	00	00
	213	0	0	1	1	00	00
	214	0	0	1	1	00	00
	210	0	0	1	1	00	00
	220	0	0	1	1	00	00
	230	0	0	1	1	00	00
	231	0	0	5	5	00	00
	232	0	0	3	3	00	00
	233	0	0	5	5	00	00
	241	0	0	1	1	00	00
	242	0	0	1	1	00	00
	243	0	0	1	1	00	00
	244	0	0	1	1	00	00
}
blahblahblash

LoginServer.ini

Code:

DB	"127.0.0.1"
Core	"127.0.0.1"

// CACHE
AddCache( "127.0.0.1" );
AddCache( "MAIN SERVER IP" );

MSG_VER	"2012120501"

//NPROTECT

CacheServer.ini

Code:

Core	"127.0.0.1"
Port	"12412"

WorldServer.ini

Code:

Sys 2
ResourceFolder	"C:\Program Files\Flyff V17 working\Resource"
Key 201
Core "127.0.0.1"
DB "127.0.0.1"
LANG	1
Heartbeat 10101
18
GUILDWAR 1
WORMON 1
DROPITEM_REMOVE
GUILDCOMBAT1TO1 1
GUILDCOMBAT 1
ARENA 1
SCHOOL 1
RECOMMEND 0
RAINBOWRACE 0
SECRETROOM 1
Proc	2
NO_TEAM_TRADE 1

If you spot anything wrong here please inform me, also i have had this working before (Under VMWare as a test server) i know i am either missing something or just being plain stupid

All ports between the 2 ip's are open (TCP AND UDP (Scope Remote Ip Address only) ) as well as all ports required to make a normal connection to the main cluster have been opened on TCP and UDP on the second cluster.

I am not using any form of server sided anti-hack to re-route ports at all (like tom's anti-hack does) so that isn't the problem.

Any help will be appreciated

Thanks
Nick

Worf · May 24, 2012

I think you need to edit the accountserver.ini to read both the IP's

ShadowDragon · May 24, 2012

SilentNL said:
I think you need to edit the accountserver.ini to read both the IP's

I don't know if he did that yet, but that is true.

NickHough · May 24, 2012

aha, yeah already done forgot to show that (thats kinda obvious if you saw the link i posted (has the 2nd cluster help but doesn't work for me) ) It keeps catching on what i assume is CacheServer as it doesn't connect to the second dedicated's ip (i checked using netstat -n in cmd prompt)

It shows
2nd Dedi IP:LOGIN PORT
1st Dedi IP

CACHE PORT)

I have also used the CoreServer for the 2nd cluster on the Main Dedicated but i can't get past the channel select if i do it like that (with the database, login, cache and world server running on dedi 2)

Also this is my AccountServer.ini from the Main cluster

Code:

TEST
AddTail( -1, 1, "Democracy", "173.233.67.186", 0, 1, 0 );
	AddTail( 1, 1, "Channel 1-1", "127.0.0.1", 0, 1, 600 );
	AddTail( 1, 2, "Channel 1-2", "127.0.0.1", 0, 1, 600 );
	AddTail( 1, 3, "Channel 1-3(PK)", "127.0.0.1", 0, 1, 600 );
AddTail( -1, 2, "Republic", "173.198.213.78", 0, 1, 0 ); // Have to admit i don't know what the ip in this exactly does ( i think it is the core servers ip? )
	AddTail( 2, 1, "Channel 2-1", "173.198.213.78", 0, 1, 600 );
	AddTail( 2, 2, "Channel 2-2", "173.198.213.78", 0, 1, 600 );
	AddTail( 2, 3, "Channel 2-3(PK)", "173.198.213.78", 0, 1, 600 );

DSN_NAME_LOGIN			"login"
DB_ADMIN_ID_LOGIN		""

DSN_NAME_LOG			"log01"
DB_ADMIN_ID_LOG			""

MSG_VER		"2012120501"

SKIP_TRACKING
DB_PWD_LOGIN	""
DB_PWD_LOG	""

I do have neuz under lock neuz (Set by Blessed source) i did however disable the account locking to make the multiple cluster possible in the accountserver.ini, do i need to disable it in the neuz as well or not?

NickHough · May 25, 2012

BUMP still need help :/

NickHough · May 28, 2012

BUMP Still frustrating me, anyone know what i've done wrong, or haven't added???

Mootie · May 28, 2012

You need a separate SQL Server setup on the 2nd computer or a remote OBDC setup(with CHARACTER_02_DBF and LOGGING_02_DBF) IIRC. It might be possible to do it another way, but that's how it's supposed to be done. Anything else is hacky and I can't help you with.

My new setup:

Serv #1(2GB RAM) - Common Services
- Account Server
- Certifier

Serv #2(12GB RAM, 2xXeon Quad 2.93ghz) - Cluster One
- Trans Server
- Core Server
- Login Server
- Cache Server
- World Server #1
- World Server #2
- World Server #3

Serv #3(12GB RAM, 2xXeon Quad 2.93ghz) - Cluster Two
- Trans Server
- Core Server
- Login Server
- Cache Server
- World Server #1
- World Server #2
- World Server #3

Serv #4(18 GB RAM, 2xXeon Quad 2.93ghz)
- [HDD Array #1] ACCOUNT_DBF
- [HDD Array #1] CHARACTER_01_DBF
- [HDD Array #1] CHARACTER_02_DBF
- [HDD Array #1] LOGGING_01_DBF
- [HDD Array #1] LOGGING_02_DBF
- [HDD Array #2] LOG_01_DBF
- [HDD Array #2] LOG_02_DBF

Each HDD Array is 2x 500GB 15K RPM. #1 is Raid 1(Redundancy), #2 is raid 0(Performance).

(btw I lol @ 64GB RAM that another private server advertises, 44GB for 2 clusters is overkill already even for officials, it's only used for the SQL Server to speed up log searching)

Anyways, you should just look at source code if you're running into issues. People did this without source code access, it should be a cake walk now that source code is public.

NickHough · Jun 7, 2012

Okay i have a new problem now, the worldserver.exe on the 2nd dedi doesn't start up properly (it gets to about 39mb memory usage then just stays there) any ideas why?

I now have it set up like this

1.AccountServer.exe
2.DatabaseServer.exe
3.CoreServer.exe
4.CoreServer Server2.exe
5.Certifier.exe
6.LoginServer.exe
7.CacheServer.exe
8.WorldServer Channel 1.exe

On the SECOND computer:

1.DatabaseServer.exe
2.loginserver.exe
3.CacheServer.exe
4.WorldServer.exe

@Lethal the source doesn't really help in this at all, i have looked there is nothing that should be blocking a addcache hook inside the LoginServer.ini but that seems to be whats happening,

For some reason the cacheserver.exe ins't hooking onto the coreserver (you know where you get the ip underneath the white line)

Mootie · Jun 8, 2012

#1 - You aren't giving much information. Run a debugger on the 2nd world and see where the exception is raised.

#2 - The source is all you ever could ask for to solve problems like this. People did this before the source was released(with PDBs). Just go through the code until you understand how the ISC(inter-server-communication) works.

NickHough · Jun 11, 2012

Lethal at comment 1, no exception is raised already done this, i am using odbc connection from the 2nd dedi to the first (secure connection only aloud between the 2 dedi ips) could this be the cause - like i said all ports are open to eachother and are scoped on the firewall to each of the dedis (both tcp and udp ports)

I can only see that the 2nd dedi's world servers aren't connecting to the core server 2 ran on the main dedi, what would be a fix for this? (as none of the 0201/0202/0203 clear their white backgrounds)

QUICK NOTE

AddTail( -1, 2, "Republic", "WHAT IS THIS IP FOR", 0, 1, 0 );
AddTail( 2, 1, "Channel 2-1", "2nd dedi ip for channel 1", 0, 1, 600 );
AddTail( 2, 2, "Channel 2-2", "2nd dedi ip for channel 2", 0, 1, 600 );
AddTail( 2, 3, "Channel 2-3(PK)", "2nd dedi ip for channel 3", 0, 1, 600 );

a print screen of the info

Core Servers - [HELP] Adding a Second Cluster (not second channel) - RaGEZONE Forums

those are after running all 3 channels on both servers (and both dedi's)) anyone have any idea what the 2nd dedi hasn't made a connection to the coreserver here (but in a error file in my program folder i get this

Code:

2012/ 6/11   04:44:40   Other Cache Connection - IP : 173.198.213.78

[HELP] Adding a Second Cluster (not second channel)

NickHough

Socius Decennalis

Worf

Master Summoner

ShadowDragon

Flyff Developer

NickHough

Socius Decennalis

NickHough

Socius Decennalis

NickHough

Socius Decennalis

Mootie

Not working on UnitedFlyf

NickHough

Socius Decennalis

Mootie

Not working on UnitedFlyf

NickHough

Socius Decennalis

RaGEZONE Sponsor

About RaGEZONE

Forum Statistics

RaGEZONE Partners