Link to home
Start Free TrialLog in
Avatar of CoolestBananas
CoolestBananas

asked on

FileSystemWatcher - reliability and huuuuuge implementation

Hi,

I am looking to implement the FileSystemWatcher in some large corporations with 5000-100000 users across as many as a few hundred data servers. I appreciate there will be bottle necks when coming to evaluate my data - I will cross that bridge later. Does anyone have experience using the FileSystemWatcher on a large scale? Is it reliable? I have read something about catching errors to ensure all activity can be acknowledged.

All thoughts will be welcome.
Avatar of Rahul Goel
Rahul Goel
Flag of India image

Avatar of CoolestBananas
CoolestBananas

ASKER

That is a very useful link. It is irritating that this class is unreliable - so predictable! Have you had any other feedback? Is this your test case?
ASKER CERTIFIED SOLUTION
Avatar of Nasir Razzaq
Nasir Razzaq
Flag of United Kingdom of Great Britain and Northern Ireland image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
I have used the filesystemwatcher to monitor files at my company (probably around 250 clients each monitoring their own network share location for new files created). I found that the biggest issue was that network interruptions meant that I had to catch the error raised and loop at intervals trying to re-initialize the filesystemwatcher until the connection was restored (otherwise the program keeps running but just desen't receive any events)


CodeCruiser: I have had a quick skim through those links and they are very useful. Thanks. I feel comfortable hooking into the Win32 API. What worries me is that this is what the FileSystemWatcher is trying to do and obviously has problems and therefore is unreliable. Why is it unreliable? Will I not just come across the same problems but in a more roundabout way?

Solar_Flare: I would be running the "watcher" on the device and building a file to be dumped on another server for evaluation. What do you think?
FileSystemWatcher is developed so that it just raises all file creation, modification, deletion events etc. When you hook into API yourself, you could handle the situations so that you do could handle the creation of a file properly. You could also code around the Watcher in way that you monitor the events being raised and match them to previous recent events to detect if multiple events are generated for same file etc.