try/catch on performance gather runs amuck..sometime s

[Greg Merideth]

I have a custom service running under win2k3 that grabs performance data
and stores it into a sql database. I've noticed on occasion when the
data gathering stops, using the sysinternals process explorer that the
open handles of the process rocket from the usual 120+ to around 3-4,000
and every few seconds the handles increase into oblivion unless I kill
the process.

Using my own xml logger, I've pinned the code that causes the run away
handles to this:

try
{
totalUCEMessage s=(int)Math.Rou nd(_perfmonT.Ne xtValue());
totalMessagesSc anned = (int)Math.Round (_perfmonT.Next Value());
}
catch(Exception exceptionError)
{
xmlLogger.AddWi ndowsEvent("Gen eral Exception " +
exceptionError. Message,_ON_STA RT_,
FwdTech.Utiliti es.Logging.XmlE ventLogger.Even tClass.Informat ion);
}

The performance values I'm attempting to read from are the microsoft
intelligent messaging filter for exchange.

Now, the system event log never gets the general exception message
posted to it so the hang-up is in the use of .NextValue(). Is there a
possible threadding issue I'm not aware of with performance gathering
use thats causing the runaway threads?

Thanks for any suggestions.

[Nicholas Paldino [.NET/C# MVP]]

Greg,

As far as I know, there is not. While I don't know of the issue, there
are actually two other pieces of code that are being run:

exceptionError. Message
FwdTech.Utiliti es.Logging.XmlE ventLogger.Even tClass.Informat ion

Are you sure either of these isn't the cause? Also, in the
AddWindowsEvent method, is there any code in there before it writes to the
event log which could be causing the handles to open?

Hope this helps.

--
- Nicholas Paldino [.NET/C# MVP]
- mv*@spam.guard. caspershouse.co m

"Greg Merideth" <be*****@forwar dtechnology.net > wrote in message
news:3a******** ************@co mcast.com...

I have a custom service running under win2k3 that grabs performance data
and stores it into a sql database. I've noticed on occasion when the data
gathering stops, using the sysinternals process explorer that the open
handles of the process rocket from the usual 120+ to around 3-4,000 and
every few seconds the handles increase into oblivion unless I kill the
process.

Using my own xml logger, I've pinned the code that causes the run away
handles to this:

try
{
totalUCEMessage s=(int)Math.Rou nd(_perfmonT.Ne xtValue());
totalMessagesSc anned = (int)Math.Round (_perfmonT.Next Value());
}
catch(Exception exceptionError)
{
xmlLogger.AddWi ndowsEvent("Gen eral Exception " +
exceptionError. Message,_ON_STA RT_,
FwdTech.Utiliti es.Logging.XmlE ventLogger.Even tClass.Informat ion);
}

The performance values I'm attempting to read from are the microsoft
intelligent messaging filter for exchange.

Now, the system event log never gets the general exception message posted
to it so the hang-up is in the use of .NextValue(). Is there a possible
threadding issue I'm not aware of with performance gathering use thats
causing the runaway threads?

Thanks for any suggestions.