ShardManager with 12000 ms, Any ideas to solve this ?

[Tazdingo]

See my shardmanager:




Server Specifications:

1 Gbps Connection
SQL 2008 R2 Standard

What can be the reason of this ?

Thanks in advance

Edit:
Some times appear this msg

Its strange because its like someone is ddosing my server but that is not the problem

[MentaL]

[Jangan]

lol... your hardware is garbage... upgrade your processor!


you need around 4-6 processors with your hardware to run the SQL 2008 db alone... maybe 2-3 more to run silkroad, and 1 more to run the machine.


you need DUAL E5620 or replace it with a DUAL QuadCore 2x8x [2.20 -> 3.12] or any high end duel processors!



dont be cheap -.-


----------------------------------

or... read my shard lag guide and disable LOG DB.... or get a 2nd machine... so many options.

[Tazdingo]

I test with alot of processors, but this one is enough for this kind of server.
Doenst matter the hardware in this case.

Is something related with the procces who manage the shardmanager, right now i disable the FTP( i have the ftp in an external server) for test purposes and reduce a little bit.
But i think is something related with the db side, something with the _TimedJob table or someting like that because i already test alot of things and nothing solve the problem.

[Jangan]

I test with alot of processors, but this one is enough for this kind of server.
Doenst matter the hardware in this case.

Is something related with the procces who manage the shardmanager, right now i disable the FTP( i have the ftp in an external server) for test purposes and reduce a little bit.
But i think is something related with the db side, something with the _TimedJob table or someting like that because i already test alot of things and nothing solve the problem.

go to university "when you finish school" and learn about how machines work, then come back and read what the fuck you just wrote. and laugh at yourself. hardware is the #1 priority to doing anything ....


there is around 500 million things i can personally do to get 0ms lag.. but, you will have to sacrifice a lot of things. if you want better performance, you need better hardware.. if you want little performance, then delete logging and 90% of other shard related things.


but DO NOT EVER !! Fucking write Hardware doesnt matter -.- ...

[Tazdingo]

go to university "when you finish school" and learn about how machines work, then come back and read what the fuck you just wrote. and laugh at yourself. hardware is the #1 priority to doing anything ....


there is around 500 million things i can personally do to get 0ms lag.. but, you will have to sacrifice a lot of things. if you want better performance, you need better hardware.. if you want little performance, then delete logging and 90% of other shard related things.


but DO NOT EVER !! Fucking write Hardware doesnt matter -.- ...

Sorry, I misspoke
What i meant with "hardware doenst matter in this case" is that i probe with better proccesors and more ram too and the problem still.
I know that the hardware is the most important thing in the computer system, first come the hardware and then the rest.
Dont blame at me for a misspoke lol xD

Can you say the most important things in that 500 millon things that you can do ? (Or just give an example)

Thanks for answer

[Jangan]

i can take a look at your database... but generally... look at your LOG_DB ... disable all the logs
http://forum.ragezone.com/f723/guide...-count-799221/


another thing you can do is lowering monster count [do the first thing, and see if it will work with you]

but yeah, generally i need to take a look at what is using most of the shardmanager resources... to tell you how to reduce it, but seriously, upgrading hardware works really well.

you might also just have data loss going from server 1 -> server 2 or backwards, if that happens and the relay time is too long, the shard will crash, which happens i guess to you.



so yeah... lots of things to verify, its probably not just 1 problem.

[Tazdingo]

i can take a look at your database... but generally... look at your LOG_DB ... disable all the logs
http://forum.ragezone.com/f723/guide...-count-799221/


another thing you can do is lowering monster count [do the first thing, and see if it will work with you]

but yeah, generally i need to take a look at what is using most of the shardmanager resources... to tell you how to reduce it, but seriously, upgrading hardware works really well.

you might also just have data loss going from server 1 -> server 2 or backwards, if that happens and the relay time is too long, the shard will crash, which happens i guess to you.



so yeah... lots of things to verify, its probably not just 1 problem.

I go to sleep now but tomorrow i will reply this post.

I will answer you some thing:
1- I disable all the logs already but the problem still (i only have enable the LogEventChar for online/offline status via website and last ip connection)
2- I already lowered the count of mobs 13,4k mobs lul, problem still
3- I have an internal uplink between my servers so the connection is pretty good, the second gameserver sometimes start after the first one.

Thanks for reply this post, maybe will help some people with the same problem :D
Tomorrow i will reply again

[MGProjects]

only a thing.

BAD PROCESSOR!
700+ char on with 2 processor? LOL need min Real processor with 4 processor for work fine 6 :)

[Nah001]

just for lulz can you make screen

Cpu usage % performance

[Tazdingo]

I remark that i test my server in alot of servers include blacklotus ones that are enough for silkroad.
In this case its not about the hardware ...

I will attach the CPU usage:


Edit:
This is my shardmanager when its start to have big ms
See the msgcount, i think that isnt normal

[MGProjects]

I remark that i test my server in alot of servers include blacklotus ones that are enough for silkroad.
In this case its not about the hardware ...

I will attach the CPU usage:


Edit:
This is my shardmanager when its start to have big ms
See the msgcount, i think that isnt normal

MsgCount very high lol.. check your internet too..

[P0kemonMast]

MsgCount very high lol.. check your internet too..

Something is odd there in those pics. I wouldn't trust that cpu usage performance if I was you. Theres no way in hell with with 700 user connected or even 600+ user your cpu is only at 3% performance. Especially since is only 2 cores cpu - _ -

[Tazdingo]

xD I'm read the replys and all is "internet conection" "hardware" etc etc etc
Is something related to the dbside or something related to configurations.

I said already that i test the same server more than 1 month in a "Dual Intel Xeon E5606" with 1gbps connection (blacklotus.net serious company) and the problem still the same.

If you dont know please dont answer "internet" "hardware" etc. Is something related to the configurations, This server have a trayectory of 1 year, this problem start with no reason three or five months ago.

Anyway thanks for try to help

Edit:
Check this images
Did you see ? Is something totally strange ! And its not hardware,internet, etc related.

[TitusPullo]

Are you using dos fixed agent server?

and

ur disc type? SATA,SAS,SSD? & have raid?

[Jangan]

lol @ the 12 GB SQL server usage!!! LOOOOL!!! you seriously need better processing power.. my god! my sql server never goes over 3-4gb of ram...

erm... yeah just like TitusPullo said... having SOLID STATE harddrive = incrases performance 1m times... since you i think the right number was 8,000:1 ratio x times better than normal hard drives :3

so try that!