🚀 go-pugleaf

mike wrote:
> I have a site that is getting 10053 errors and I am not sure how to handle 
> this in my application.

You did a search for what 10053 stands for ? Please supply the description
also.

> I have a central host that allows many clients to connect in a blocking 
> socket manner, 

Are you using raw sockets or an Indy component or something like that ?

Hans.

Hans,

Thanks for answering. Yes the message description of the error is as 
follows:
WSAECONNABORTED (10053) Software caused connection abort.

Berkeley description: A connection abort was caused internal to your host 
machine. The software caused a connection abort because there is no space on 
the socket's queue and the socket cannot receive further connections.

WinSock description: Partly the same as Berkeley. The error can occur when 
the local network system aborts a connection. This would occur if WinSock 
aborts an established connection after data retransmission fails (receiver 
never acknowledges data sent on a datastream socket).

TCP/IP scenario: A connection will timeout if the local system doesn't 
receive an (ACK)nowledgement for data sent. It would also timeout if a 
(FIN)ish TCP packet is not ACK'd (and even if the FIN is ACK'd, it will 
eventually timeout if a FIN is not returned).

I am using the TSocketServer from builder not Indy or any anything else.



Mike



"Hans Galema" <notused@notused.nl> wrote in message 
news:42c5783a$1@newsgroups.borland.com...
> mike wrote:
>> I have a site that is getting 10053 errors and I am not sure how to 
>> handle this in my application.
>
> You did a search for what 10053 stands for ? Please supply the description
> also.
>
>> I have a central host that allows many clients to connect in a blocking 
>> socket manner,
>
> Are you using raw sockets or an Indy component or something like that ?
>
> Hans. 

mike wrote:

> I have a central host that allows many clients to connect in a blocking 
> socket manner, so I am creating a thread for each active client connection. 
> So what happens is all works well for a number of hours then the 10053 
> errors start.

How many connections are there usually whan that happens ? How do
the clients behave: do they come and go within a few minutes or
are they staying for hours ?

> This of course breaks the socket connection to the client and 
> then the client attempts to connect again. When the client connects the 
> 10053 errors persist and clients cannot connect.

So such a client can reconnect but then after a short time there is
the same error ? Is this for some specific clients or for all then?

What about clients that did not connect before the errors stated. Are
they handled ok ?

> So can someone assist me in understanding why this error would start 
> occurring, and then once they start how do I get my application back into 
> the state where connections can resume.

Aha. So a client can not resume ? Please explain in detail.

Did you do test with all clients and the server on the same machine ?

Hans.

Hans,

Thank you for replying, I appreciate it very much. This problem is really 
causing my hair to turn grey and I really need to figure it out.


> How many connections are there usually whan that happens ? How do
> the clients behave: do they come and go within a few minutes or
> are they staying for hours ?

Well I can't really determine how many connections there are when the 
problem happens but I at any given point I could have 40 to 80+ machines 
connected.

Once connected the clients stay connected for hours at a time.

> So such a client can reconnect but then after a short time there is
> the same error ? Is this for some specific clients or for all then?

Yes, once the problem starts a client may connect but in a very short time 
the error occurs.
Also once the error starts all the previos clients start suffering the 
problem, then they become dissconnected and now a really bad cycle starts as 
they all begin to reconnect and the error occurs which means they disconnect 
again and the cycle starts.

It effects all the clients, not a specific client.


>
> What about clients that did not connect before the errors stated. Are
> they handled ok ?
>

No, once the problem starts all clients are effected, new clients, old 
clients, all clients.

Its like the Host Socket has gone ito some kind of error and it effects all 
clients, both connecting and those previously connected.

> Aha. So a client can not resume ? Please explain in detail.

Well what happens is the client detects the socket disconnect, and, once 
this happens the client attempts to reestablish connection to the host. The 
gets a connection, and a couple of messages are transmitted back and forth 
and then the error happens, the client disconnects, the cycle starts over 
again.



>
> Did you do test with all clients and the server on the same machine ?

Ummmm, no, as this will be a very difficult thing to do as the problem is 
happening at a client site.

I do have the same software running at other sites without any of the above 
problems. But what I don't understand is why it could possibly be happening, 
and also, once it starts to happen what can I do with the host application 
to attempt to correct the situation.

Again, thanks Hans for trying to help me with this nasty problem.

Mike





"Hans Galema" <notused@notused.nl> wrote in message 
news:42c653f0@newsgroups.borland.com...
> mike wrote:
>
>> I have a central host that allows many clients to connect in a blocking 
>> socket manner, so I am creating a thread for each active client 
>> connection. So what happens is all works well for a number of hours then 
>> the 10053 errors start.
>
> How many connections are there usually whan that happens ? How do
> the clients behave: do they come and go within a few minutes or
> are they staying for hours ?
>
>> This of course breaks the socket connection to the client and then the 
>> client attempts to connect again. When the client connects the 10053 
>> errors persist and clients cannot connect.
>
> So such a client can reconnect but then after a short time there is
> the same error ? Is this for some specific clients or for all then?
>
> What about clients that did not connect before the errors stated. Are
> they handled ok ?
>
>> So can someone assist me in understanding why this error would start 
>> occurring, and then once they start how do I get my application back into 
>> the state where connections can resume.
>
> Aha. So a client can not resume ? Please explain in detail.
>
> Did you do test with all clients and the server on the same machine ?
>
> Hans. 

mike wrote:

> I do have the same software running at other sites without any of the above 
> problems.

Then it is very difficult to debug indeed. Is the OS different there ?
Or has the server much less RAM ? You should really count
your serverthreads so you know at which amount problems begin.

> But what I don't understand is why it could possibly be happening, 
> and also, once it starts to happen what can I do with the host application 
> to attempt to correct the situation.

Did you try to delete all threads, Close() the TServerSocket and then
to Open() it again ?

Please have a look at the way you reply. It makes no sence to reply
selective and also totallly.

Hans.

Hans,

Thanks for replying.

>Then it is very difficult to debug indeed. Is the OS different there ?
> Or has the server much less RAM ? You should really count
> your serverthreads so you know at which amount problems begin.

Yes this is hard to debug.  The server is running 2000server and it does not 
have less ram then other sites where all work fine.

I will make a change to the server software to count the threads that are 
open when the first 10053 error occurs and place that in a log file. At this 
point I am not sure what the count is.

> Did you try to delete all threads, Close() the TServerSocket and then
> to Open() it again ?

No I have not tried that yet, but will modify the server software to do this 
at the first 10053 received to see if this clears it up.

If I understand the error message description properly, the server is 
sending a TCP message and not getting back the ACK from the client. Is it 
possible that a router somewhere in the mix could be causing this to happen. 
If so do you have any insight as to what I could ask the site to look for on 
the router. I am not a router guy so I am at a bit of a loss on that.

It may take me a few days to get the software modified and installed at the 
site to test.

Boy this is troublesome.

Thanks for the assitance.

Mike



"Hans Galema" <notused@notused.nl> wrote in message 
news:42c983b3$1@newsgroups.borland.com...
> mike wrote:
>
>> I do have the same software running at other sites without any of the 
>> above problems.
>
> Then it is very difficult to debug indeed. Is the OS different there ?
> Or has the server much less RAM ? You should really count
> your serverthreads so you know at which amount problems begin.
>
>> But what I don't understand is why it could possibly be happening, and 
>> also, once it starts to happen what can I do with the host application to 
>> attempt to correct the situation.
>
> Did you try to delete all threads, Close() the TServerSocket and then
> to Open() it again ?
>
> Please have a look at the way you reply. It makes no sence to reply
> selective and also totallly.
>
> Hans. 

mike wrote:

> If I understand the error message description properly, the server is 
> sending a TCP message and not getting back the ACK from the client. Is it 
> possible that a router somewhere in the mix could be causing this to happen. 

Sorry, but that is beyond my knowledge. Could the errors be
caused by one clientconnection only ? Is that what you are suggesting ?
Not the serverrouter but a router somewhere in a cliets route ? Why is
it only after hours ?

>>Please have a look at the way you reply. It makes no sence to reply
>>selective and also totallly.

Apparantly you overlooked that. Please do not quote a complete message.
And certainly not if you already quoted selective.

Hans.

Hans Galema wrote:

> Sorry, but that is beyond my knowledge. Could the errors be
> caused by one clientconnection only ? 

If you log the 10053 error then you log the client-ip too ?
See if it is always the same one.

Hans.

Hans,

> Sorry, but that is beyond my knowledge. Could the errors be
> caused by one clientconnection only ? Is that what you are suggesting ?
> Not the serverrouter but a router somewhere in a cliets route ? Why is
> it only after hours ?

No the problem happens during the day. The site is closed at night. The 
reason I was wondering about a router or something like that is because I am 
trying to explore all possibilities, and routers and such are certainly in 
the mix.


Sorry about the replies, I have corrected my actions.

Mike 

Hans,

Yes I will log the IP as well. At this point I don't believe it is a single 
client starting the problem, but I may be wrong on this.

Mike



"Hans Galema" <notused@notused.nl> wrote in message 
news:42c99656@newsgroups.borland.com...
> Hans Galema wrote:
>
>> Sorry, but that is beyond my knowledge. Could the errors be
>> caused by one clientconnection only ?
>
> If you log the 10053 error then you log the client-ip too ?
> See if it is always the same one.
>
> Hans. 

First I want to thank Hans for taking the time to respond to my problem, I 
apprciate it very much.

After many, many hours of studing and tracing the problem I have determined 
that it is not a software problem but a hardware problem within the network. 
It turned out to be a switch on the network which had gone terribly bad and 
was causing packets to be dropped, bad crc's, and just generally wrecking 
hovac on the network.

So the 10053 error was an accurate error as far as the software was 
concerned, I just didn't believe it and immediately suspected my software.

Again thanks for responding, it helped me zero in on the problem.

Mike

"Hans Galema" <notused@notused.nl> wrote in message 
news:42c99656@newsgroups.borland.com...
> Hans Galema wrote:
>
>> Sorry, but that is beyond my knowledge. Could the errors be
>> caused by one clientconnection only ?
>
> If you log the 10053 error then you log the client-ip too ?
> See if it is always the same one.
>
> Hans. 

mike wrote:

> After many, many hours of studing and tracing the problem I have determined 
> that it is not a software problem but a hardware problem within the network. 
> It turned out to be a switch on the network which had gone terribly bad and 
> was causing packets to be dropped, bad crc's, and just generally wrecking 
> hovac on the network.
> 
> So the 10053 error was an accurate error as far as the software was 
> concerned, I just didn't believe it and immediately suspected my software.
> 
> Again thanks for responding, it helped me zero in on the problem.

And thank you for informing us of this. It is good to know that all ended well.

Hans.