> The linchpin of all this is, that the program itself is up and 
> running, doesn't hang and that it can
> still response. Otherwise, days can go by and no data is recorded.

I hope your application is already a Windows service, you can then make use of
the various service restart options if it stops. 

These options do not handle lock-ups, so you really need a second Windows
service sending messages to the first, and using various methods to crash and
restart it if there is no response, maybe also checking a shared file or
registry key to make sure the first program is working.  

> So my idea was to have a little watchdog (to be programmed) running 
> on a server which regularly connects to all of the clients to get a
> response that all is well. 

Several years ago I had a large project with PCs running on road vehicles,
which is a horrible environment, power surges, heat and vibration.  I had three
'vehicle activity servers' using TWSocketServer which the remote vehicles
contacted via wifi (and later 3G) every five minutes, which updated a SQL
database, reported on web pages.  

I designed a simple TCP/IP ASCII protocol that ended up with about 50 different
commands each way, returning status information like fan speeds and CPU
temperatures so we knew when the PCs were about to die (50% of CPU fans died
within a year).  

Some of this you can do with off the shelf software, I sell an application
ComCap that accepts UDP and TCP messages, and will email based on content or
when they stop, and I use it myself so I know when my remote hosted servers
goes off-line.  

Nagios is probably similar, but much more specialised than ComCap, but I've no
idea what sort of input streams it expects. 


To unsubscribe or change your settings for TWSocket mailing list
please goto http://lists.elists.org/cgi-bin/mailman/listinfo/twsocket
Visit our website at http://www.overbyte.be

Reply via email to