[ENet-discuss] Auto disconnect at client witheventENET_EVENT_TYPE_DISCONNECT

Tibor Klajnscek tibor.klajnscek at gmail.com
Mon Sep 19 13:53:57 PDT 2005


You should check a few things:

1) Make sure that the low-level socket functions of enet are getting called
2) If they are not, than something's terribly fubar (in our game that 
uses enet, we used to do a whole level load
    which took up to 30 seconds and we managed to keep the connection 
alive without sending anything since
    the loading code is blocking )
3) If they are, check for socket errors - are you sure you are not doing 
anything funky with the sockets?


Lee Salzman wrote:

> In theory, those are the only things that effect how long it takes to 
> timeout.
>
> Lee
>
> Thomas Johansson wrote:
>
>> Im having trouble understanding how this works.
>>  
>> I have a program where a client connects to a server. Both sides 
>> sends only reliable messages. This works fine at first.
>>  
>> Then my client preforms an operation that take several seconds, this 
>> causes a timeout due to the fact that no enet_host_service occurs.
>>  
>> I then try to increse the timeout limits by changing to the following:
>> NET_PEER_TIMEOUT_LIMIT                = 128, 
>> ENET_PEER_TIMEOUT_MINIMUM              = 10000,
>> ENET_PEER_TIMEOUT_MAXIMUM              = 50000,
>>  
>> Both on client and server.
>>  
>> After I do this, no timeouts occur. However no messages I send seem 
>> to reach the server after the point in time where the timeout 
>> previously occured and nothing gets acked (thus a timeout occurs 
>> eventually).
>>  
>> What could it be? Im confused. Are there additional timeout limits I 
>> need to set or am I messing up the throttling algorithm by
>> changing the timeout values?
>> /Thomas
>>
>>     ----- Original Message -----
>>     *From:* Thomas Johansson <mailto:thomas at paradoxplaza.com>
>>     *To:* Discussion of the ENet library <mailto:enet-discuss at cubik.org>
>>     *Sent:* Friday, September 16, 2005 10:50 AM
>>     *Subject:* Re: [ENet-discuss] Auto disconnect at client
>>     witheventENET_EVENT_TYPE_DISCONNECT
>>
>>     No takers on this? Im also very interested in these issues.
>>     /Thomas
>>
>>         ----- Original Message -----
>>         *From:* Ulhas Dhuri. <mailto:ulhas.dhuri at gmail.com>
>>         *To:* enet-discuss at cubik.org <mailto:enet-discuss at cubik.org>
>>         *Sent:* Thursday, September 08, 2005 11:31 AM
>>         *Subject:* [ENet-discuss] Auto disconnect at client with
>>         eventENET_EVENT_TYPE_DISCONNECT
>>
>>         Hi!
>>
>>         Looking for any active enetian, who could help out :),
>>
>>         I was doing some stress testing for enet library on my LAN .
>>
>>         1> I use Enet Server to to accept 2000 clients and send them
>>         small test data and also receive small
>>               test data. (In Infinte loop ) at interval of 1 second.
>>
>>         2> I use  Enet multithreaded  program which spawns  300 test
>>         clients sending/receiving  some test data at interval of 1 
>> second .
>>
>>
>>                                    Now it happens that the slowly
>>         client's  connection breaks down posting event
>>         ENET_EVENT_TYPE_DISCONNECT , where as the server didn't
>>         disconnect the client .
>>
>>
>>         On a Lan under ideal circumstances, there should be no such
>>         TIMEOUT .I further tried to explore into enet code and did the
>>         following stuff:
>>
>>          1>set in enet.h                  
>> ENET_PEER_TIMEOUT_LIMIT                = 64
>>
>>          2>protocol.c line no :913 ,       if (outgoingCommand ->
>>         roundTripTimeout >= (outgoingCommand -> roundTripTimeoutLimit 
>> )*400)
>>          ie. increased the roundTripTimeoutLimit by 400 times
>>
>>         3>tried out
>>         enet_peer_throttle_configure(*tempIter,5000,10,500);
>>
>>         but all in vain .
>>
>>
>>         Could some one with a good understanding on enet post some
>>         solution to this problem and throw some more light on
>>         enet_peer_throttle_configure() making it usage more clear
>>
>>
>>         --         Warm Regards,
>>                                     Ulhas . S .  Dhuri
>>            
>>         
>> ------------------------------------------------------------------------
>>
>>         _______________________________________________
>>         ENet-discuss mailing list
>>         ENet-discuss at cubik.org
>>         http://lists.cubik.org/mailman/listinfo/enet-discuss
>>
>>     
>> ------------------------------------------------------------------------
>>
>>     _______________________________________________
>>     ENet-discuss mailing list
>>     ENet-discuss at cubik.org
>>     http://lists.cubik.org/mailman/listinfo/enet-discuss
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> ENet-discuss mailing list
>> ENet-discuss at cubik.org
>> http://lists.cubik.org/mailman/listinfo/enet-discuss
>
>
> _______________________________________________
> ENet-discuss mailing list
> ENet-discuss at cubik.org
> http://lists.cubik.org/mailman/listinfo/enet-discuss
>



-- 
No virus found in this outgoing message.
Checked by AVG Anti-Virus.
Version: 7.0.344 / Virus Database: 267.11.1/104 - Release Date: 9/16/2005



More information about the ENet-discuss mailing list