in reply to IO::Socket Not responding after period of time and traffic!

Probably not the cause of your cited problem, but I'm utterly astounded that your code works at all. Given that the 6 lines in the second block below are missing the keyword substr? 

                $day=substr($_,42,2);
                $month=substr($_,44,2);
                $year=substr($_,46,2);

                $epanic=($_,49,1);
                $eoverspeed=($_,52,1);
                $egeoin=($_,52,1);
                $egeoout=($_,52,1);
                $egeooutspeed=($_,52,1);
                $egeoinspeed=($_,52,1);

                $event=substr($_,48,5);

[download]

I guess either errors are rare, or not handling them means ignorance is bliss. I have to award you the prize for some of the nastiest code I've seen this year.

Had you used strict and warnings, you might have noticed that. Along with the typo in:

            print "Socket Created On : $sockeaddress:$socketport\n\n";

[download]
and the bad escape in $data = "!$msgreply\OK";.

With those, judicious use of unpack, and a little formatting and you could end up with 120 lines that looks like this rather than the 260 lines of yuck you posted. And the monks might be more tempted to look closely.

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
"Science is about questioning the status quo. Questioning authority".
In the absence of evidence, opinion is indistinguishable from prejudice.
RIP an inspiration; A true Folk's Guy

Replies are listed 'Best First'.
Re^2: IO::Socket Not responding after period of time and traffic!
by aaronwroblewski (Initiate) on Aug 16, 2010 at 09:16 UTC

    I broke out in laughter when I read this, I have a PostIT note here to look at why some data was not ending up in the DB (haven't looked at it yet) and you solved it for me by pointing out a simple oversight. Thanks.

    Thanks for the other hints of encoragement, it nice to know I have the award for something prestiges lol.

    How did you tidy up my code so easy and well, I am sure you did not just rewrite it?!.

    Thanks again!
[reply]
      How did you tidy up my code so easy and well, I am sure you did not just rewrite it?!.

      Sorry to disappoint, but I spent about 15 minutes with my editor. A couple of macros for in/undenting, but mostly just by hand. It's a simple, almost mechanical process that I apply to most every piece of badly structured code I look at in detail. It's second nature, and helps me understand the code.

      The most complex task was converting your substrs to unpack. I marked all the lines containing "substr", copied them to another buffer, sorted them into order and did a little block c&p to create the variable list and template. That's when I discovered the missing keywords.

      For solving your described problem. The first thing I would seek to eliminate, is any bad interaction with the DB.

      I'd set a version going that instead of performing the actual DB calls, have it just write the formatted SQL to a file. That will allow you to avoid losing the data because it will be a simple process to use your DBs command line client to apply the SQL manually.

      The particular possibility I'm thinking of here is that client is running your DB server out of handles. A few years ago I encountered a similar problem with MySql where the server was configured with a relatively low number of concurrent handles, and a relatively long timeout (~900 seconds). This meant that after a period of particulary high activity, inbound connections to the DB would block awaiting a handle to timeout and become available.

      The solution--beyond upping the number available and trimming back the timeout--was to ensure that all statement handles were correctly finished; and all DB handles were correctly disconnected.

      You are doing the latter, but not the former, and it might be the cause here. The nice thing is that by writing the SQL to a file, you can quickly find out if it is without data loss.

      Anyway, if not, the tidied up code should be easier to debug. You might (for example) consider using Devel::Trace to determine where in the code it is hanging.

      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      "Science is about questioning the status quo. Questioning authority".
      In the absence of evidence, opinion is indistinguishable from prejudice.
[reply]

In Section Seekers of Perl Wisdom