Problem with "sqlist_check_timeout"

It does not have a pattern.

The symptom of this "anomaly" is the blocking of port 6112 in the selection of chars..

for now I have solved the problem by reducing the time of empty sockets checking on the server

# How often the server will check queues for expired data
sq_checkinterval =

# How long will it take for a server to expire?
sq_timeout =

but I've seen this on other forums, it seems to be an attack with "ghost connections" (empty connections). They serve only to occupy all sockets available for connection on the server.

I think that's it..

Added: 04.10.2017 16:04

Our friend Harpywar could help us develop something in D2CS to avoid these "ghost connections."

Added: 04.10.2017 17:14

don't work...

I will try to create a rule in iptables to limit the connections on the ports due.. return if it succees!

Added: 06.10.2017 09:48

dont work!

Added: 07.10.2017 10:54

I'm thinking it might be something with OS.

Some kernel errors appeared ... Immediately put here.

Could you help me HarpyWar?

bump

not yet...

latest version of Pvpgn

It seems that if you have a time limit or use.. every 2 days connected this fault occurs.

I'm going to start renaming D2CS every 24 hours to see if the problem goes away, if at all, let's know that it is related to some kind of failure/limit in OS.

If it were the rules of the firewall, why would it work to a certain extent?

I have a few IPtables rules and the problem started before adding the rules

When I have a while I will try to use another version of PVPgn or another version of OS.

But I would like more options, to make the changes, I did not want to have to keep stopping the server to perform these maintenance: /

What would be the most stable version to use on the Linux system?

The latest version "PvPGN 1.99.7.1-PRO"?

or

The "PvPGN v1.99-SVN" version?

I ask this question because I use the latest version of PvPGN in my OS Linux and I'm having problems with "sqlist".

This problem -> https://forums.pvpgn.pro/viewtopic.php?id=1842

It looks like it does not clean or refresh this data and when it reaches the 32,000 mark it hangs from D2CS.

I have to be rebooting the D2CS every 2 days due to the volume of players

If one day increase the volume of players, I will have to decrease the restart time to 1 days or less.

Tnks

Attach full logs bnetd, d2cs, d2dbs, d2gs with maximum loglevel.

(I recommend use the latest pvpgn from Github)

Help me @HarpyWar

The problem returned, in another S.O (Debian 8) and with the last version of PvPGN PRO.

Identical to the above problem, could you point me in a direction of what it might be?

Failed D2CS? Because? What would this "sqlist" be? What reasons for this queue to be full?

Some direction to search for @HarpyWar please.

Thank you

Added: 02.02.2018 15:41

When I tried to use the /rehash all command to reload the D2CS these errors occurred to me:

Feb 02 11:59:45 [error] net_listen: error bind listen socket
Feb 02 11:59:45 [error] server_listen: error listen socket
Feb 02 11:59:45 [error] d2cs_server_process: failed to setup listen socket
Feb 02 11:59:45 [error] main: failed to run server

Added: 05.02.2018 07:40

Update: I am now using the PvPGN PRO Windows version running with Wine on Debian 8 64-bit OS.

I am using the same configuration in PvPGN, both in the Linux/Windowns version.

Apparently, the problem was in PvPGN for Linux, maybe something in the build, or in my OS.

So far working well, keep you informed.

Added: 06.02.2018 12:19

dont work, same problem