Trinitycore: StartNetwork failed to bind socket acceptor

[ghost]

hi guys

i use in Linux from autorestart for worldserver

A new problem has happened

When I restart the server with the following command
.server restart 60
but .. server failed to full start work

Error:

World initialized in 0 minutes 8 seconds
StartNetwork failed to bind socket acceptor
Failed to initialize network
Couldn't bind to 0.0.0.0:7878

/home/TC/Core/src/server/shared/Networking/SocketMgr.h:35 in ~SocketMgr ASSERTION FAILED:
!_threads && !_acceptor && !_threadCount StopNetwork must be called prior to SocketMgr destruction

Segmentation fault (core dumped)

Before that, there was not a problem with the new update

There is a server on realm!

To resolve the problem, I need to reset the entire Linux server

what is the problem

TRINITYCORE 3.3.5 LAST VERSION CLONER > DOWNLOADED.

OS: Linux debian

[TrinityCoreBot]

This issue is not considered valid because of the following reasons:

• No proper commit hash (revision)

Please read http://bit.ly/tc-issuetracker-and-you and add the missing information to this issue. Thanks.

[jackpoz]

which autorestarter are you using ?

[ghost]

hi

rerestart_check.sh
restart.sh

chmod a+x restart.sh && chmod a+x restart_check.sh

./restart.sh

=======================

There was no problem in this new update!

[ghost]

Yeah, it's related to this https://github.com/TrinityCore/TrinityCore/commit/7874bee7bfb70e0e039f91173cff212e9572de09

I've experienced the same issue with worldserver, but strangely it just freezes instead of shutting down (and the freeze detector is not able to hang it from that state either).

Authserver also keeps crashing until the port is freed if you close it with ctrl+c.

I however thought it was an expected behavior to wait for the OS to free the ports if the application shuts down abruptly.

Running on Debian

[ghost]

y
It's been so bad

but I like it shauren :D

[jackpoz]

just wait a few seconds before restarting the application, adding some kind of wait in your script

[Shauren]

Few seconds by default is 2 or 4 minutes on linux

[ghost]

Wait a little longer than 4 minutes to get the server up?
I've been testing for more than 6 minutes, but it's not up to you to remember why it should be
Return to the first state
This is not good.
............

[Shauren]

I say this is good. At least you don't get people coming here complaining that they started the server and can't connect because there is no info that port was taken

[GreatRedem]

@Shauren, they needs to keep authserver restart every min to keep it Up (easy way)

authserver easily get stuck by flooding packet, and this cannot be fixed by firewalls, because connection has ESTABLISHED

Reproduct: Get connect to authserver then flood _HandleRealmList and you are done

[Shauren]

Why are you bloody idiots complaining that autorestarting doesnt work instead of reporting the main issue which is realmlist flooding? (and that issue is something I would work on)

[ghost]

"The problem is when you have to wait a few minutes to work؟!"

.server restart
soap connect is enable for website.
I even waited 10 minutes, but did not work : error worldserver for soap port 7878 is ready use!

World initialized in 0 minutes 8 seconds
StartNetwork failed to bind socket acceptor
Failed to initialize network
Couldn't bind to 0.0.0.0:7878

and

There is an application that flood the authserver = stopped freeze authserver

Which is solved by the problem autorestarting
"if fast start...."

Did you understand؟

You say is it right?

It was not a problem before
Remove this new update.

[Shauren]

FYI - soap port configuration was not touched.

[jackpoz]

Remove this new update.

TC is opensource, you can edit the source code to remove the update on your local checkout/own fork. Restarting authserver every minute is not TC solution to fix any issue so you should ask to who provided you that solution.

As for how to properly unbind the port on shutdown so that worldserver/authserver can be restarted right away, that's another thing and worth to be checked.

[jackpoz]

Throwing in a random word "SO_REUSEADDR"

[ghost]

tested and post now edit.
if worldserver Restart Or Crash!

A Realm is only active

.server restart
if soap connect is enable for website.
I even waited 10 minutes, but did not work : error worldserver for soap port 7878 is ready use!

World initialized in 0 minutes 8 seconds
StartNetwork failed to bind socket acceptor
Failed to initialize network
Couldn't bind to 0.0.0.0:7878

/TC/server/shared/Networking/SocketMgr.h:35 in ~SocketMgr ASSERTION FAILED:
!_threads && !_acceptor && !_threadCount StopNetwork must be called prior to SocketMgr destruction
Segmentation fault (core dumped)

OS: ON All of LiNUX

[nawuko]

Can we just close this? Either open a new issue that the port isn't correctly unbound, or create a issue with a feature request to handle authserver floods better (for example the authserver dosent have a timeout currently etc). But spamming the issue template multiple times will not help TC nor you in any way.

[shenhuyong]

@irgc Try revert commit 7874bee7bfb70e0e039f91173cff212e9572de09 and test it.

[Aokromes]

funny thing, still no revision after 5 days.

[ghost]

now tested - ON LiNUX

Authserver
authserver closed by cntrl + c and kill or more close..

Start authserver is Fast but! It takes 2 minutes to enter the server.
server is down. waiting players for 2min and login on server!

not good if hacker use flood on authserver..
There is such a program that does not resolve the firewall

If the login to the server is fast , Can be solved by authrestart !
hacker flood authserver = freeze stopped . but we can set authserver on authrestart
if hacker use flood attacks on authserver for close and stop login players on server!
authrestart can be started fast and players login on server!
So I said there is a difficulty with the new update

&
And the more important problem
&

worldserver
on worldserver.cfg enable soap 7878 connects.. for website
if "worldserver" restarted and stopped , To reboot the problem
.server restart
Not running Again

Error:

StartNetwork failed to bind socket acceptor
Failed to initialize network
Couldn't bind to 0.0.0.0:7878
terminate called without an active exception
Segmentation fault (core dumped)

Freeze worldserver.

To solve the problem, only the rest of the Linux system has to be reset

This problem is not just for me, it is for everyone

This update is problematic for TrinityCore
https://github.com/TrinityCore/TrinityCore/commit/7874bee7bfb70e0e039f91173cff212e9572de09

[jackpoz]

Btw you know, if a hacker is targeting your authserver, you can restart it how many times you want, the hacker will target again authserver so it will freeze it again. Did you setup a firewall rule about how many connections a source ip can open/close every few seconds or so ?

I'd also suggest you to take a couple of minutes to write a more readable comment. There's also no need to write the same things over and over.

Before the update you mentioned SOAP binding would most likely fail to bind on restart which was an issue too.

[ghost]

@jackpoz my dear

On closed and restart worldserver , We had no problem before this update!
https://github.com/TrinityCore/TrinityCore/commit/7874bee7bfb70e0e039f91173cff212e9572de09

Why don't you believe!

Can not solve this problem؟

[ghost]

What should I do to help resolve this problem?

worldserver
on worldserver.cfg enable soap 7878 connects.. for website
if "worldserver" restarted and stopped , To reboot the problem
.server restart
Not running Again

Error:

StartNetwork failed to bind socket acceptor
Failed to initialize network
Couldn't bind to 0.0.0.0:7878
terminate called without an active exception
Segmentation fault (core dumped)

Freeze worldserver.

To solve the problem, only the rest of the Linux system.

This update is problematic for TrinityCore
7874bee

:disappointed:

[ghost]

@Shauren
@jackpoz

Does not support someone.

gg

[ghost]

Who is capable of solving this problem

What should I do? i need enable to soap!

When will this problem be solved..........

Ay!
:disappointed:

[Aokromes]

you still don't have supplied what revision are you using.

[ghost]

@Aokromes my deaaaar
Finally, should this problem be solved?

Why does not it work to fix this problem?

I'm tired