There are two problems with this approach. A check via a URL affects the
stats generated by IIS for the site being checked. If checking a
customers site then this can skew their stats reporting unless the
checking machines IP address is blocked from the generated stats. 

Another affect is that restarting w3svc doesn't always bring back ASP
and can cause HTTP/1.1 Application errors on busy servers.

I find it more reliable to monitor the w3svc service then undertake
iisreset via a command line. This is always reliable in restarting
services although one has to be careful that IIS MMC is shutdown
properly before hand or you could loose IIS settings for recently set up
sites (IIS MMC doesn't write setting changes to the metabase in real
time, it does this when it feels like it or when IIS MMC is shut down
gracefully).

As a back up it IS worth checking w3svc via a URL but on a longer
timescan. For instance we scan the w3svc service every minute and HTTP
via a URL (to an ASP page) every 15 minutes.

David



-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Brett Hanson
Sent: 07 March 2003 14:39
To: [EMAIL PROTECTED]
Subject: RE: [SA-list] Monitoring NT service: w3svc


One thing I've done for testing in Servers Alive is to take the user's
perspective whereever possible.  For example, for the w3svc service, I
do not do a NT service check for it.  I have Servers Alive do a URL
check to an ASP page that tests ASP's functionality.  If the ASP page
comes back OK, I assume that the Service is running properly.  If the
page does not come back within the timeout correctly, I restart the
w3svc service on the machine.  If it fails twice in a row, I am
notified.

Brett Hanson
Agrium

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] 
Behalf Of Navarrete, Jose
Sent: Thu Mar 06 11:36 PM
To: Salive (E-mail)
Subject: [SA-list] Monitoring NT service: w3svc


Hello,
I am having problems monitoring the W3SVC service in three very busy IIS
servers. Several times the Alert generates DOWN messages even thou the
Server is running fine. I had increased the timeout value to 4 seconds
to prevent the "false" alerts  but I am not sure if it is a good value
(maybe it is too
high) to have a true reading for the W3SVC service.

Any help is appreciated,
Jose.


To unsubscribe from a list, send a mail message to [EMAIL PROTECTED]

With the following in the body of the message:
   unsubscribe SAlive



To unsubscribe from a list, send a mail message to [EMAIL PROTECTED]

With the following in the body of the message:
   unsubscribe SAlive

To unsubscribe from a list, send a mail message to [EMAIL PROTECTED]
With the following in the body of the message:
   unsubscribe SAlive

To unsubscribe from a list, send a mail message to [EMAIL PROTECTED]
With the following in the body of the message:
   unsubscribe SAlive

Reply via email to