SOAP::LITE dying sometime without proper message

Our environment is Teamsite 6.1. I am trying to figure out the reason for occasional dying of SOAP::LITE module. The SOAP::LITE module is being used by us to deploy XML data to the database via a webserver. The webservice accepts XML as a string, extracts and deploys the data to the database based on the content type provided in the XML. The code works fine almost all the time. Occasionally, however I have noticed a failure at the time of calling the webmethod and the LOGGER object get destroyed without going in the expection handling piece. I am looking for a better error handling so that it won't give a false impression to the user that the deployment was successfull when in actuall it failed.

my $xmlString = SOAP: Smiley Very Happy

ata->name($WEBSERVICE_VAR)->value($xmlAsString)->type('string');

#print $xmlString."\n";
my $client = SOAP::Lite
->uri($URI)
->on_action( sub { join '', @_ } )
->proxy($webServiceaddress);

my $result;
#calling webmethods
if ($webMethodType eq "Insert")
{
$result = $client->InsertUpdate($xmlString); #----------- DIES write here
}
elsif ($webMethodType eq "Delete")
{
$result = $client->Delete($xmlString);
}
$Logger->writeMessage("\$result =". $result); #------- DOESN'T COME TO THIS LINE IF FAILS

if ($result->fault)
{
# error handling
$Logger->writeMessage("Fault String=". $result->faultstring);
return $result->faultstring;
}

Find more posts tagged with

Comments

chuckles

eval {
# do SOAP call
};
if ($@) {
print "SOAP call failed. Reason [$@]\n";
} else {
print "No error.\n";
}

sabih

Thanks for you reply. Yes that is one way to handle that. Is there anyother way other than EVAL?

chuckles

Is this a command-line tool, CGI, workflow external task? If you want to capture any errors/warnings from a script that is not a command-line, you can just create an ad hoc logger that captures everything that goes to STDERR like this

open STDERR, ">>/path/to/error.log";

You can print custom statements at various points and it will also capture and system-generated errors and warnings. Just do a tail -f on the log file when it runs (assuming you are on Unix).

sabih

Its a workflow task.

chuckles

Did you find anything out useing eval {} or capturing STDERR? We will need more info to help you out. One more tip is that you can add a few more parameters to your SOAP client. Try setting on_debug and on_fault to be subs that print to STDOUT or a log file and then run the script and see what you get. Like this:

my $debug_sub = sub {print "DEBUG: @_\n"; }
my $fault_sub = sub {print "FAULT: @_\n";}
$soap = SOAP::Lite
-> uri($uri)
-> proxy($proxy)
-> on_action( sub { sprintf '""', @_ } )
-> on_debug( $debug_sub )
-> on_fault( $fault_sub )
-> encoding('UTF-8');
;

sabih

The problem is that that issue can't always be recreated. It happend once and we don't know when it can happen again but we are trying to device a method that will not give the user a false impression of deployment being successfull in case of such failures. Thanks