Connection Timeout problem when Connecting

[SpectralForce]

I am running IIS6 (Windows 2003 Server with SP1) and have FrontPage extensions installed and subwebs set up for each department off the main web. If the server is restarted everyone seems to be able to publish to the server (using FrontPage 2003) ok. But after awhile people start getting errors saying “Error retrieving “index.html” – the connection was timed out or was aborted by the server.”.

At first I restarted IIS or just restated the server and it would work ok, but now it seems to be happening with greater occurrence. It will still bring up the folder list with the files in their Subweb, but it won’t open them. And now when it happens in a Subweb, it blocks me from being able to open the main index.htm file in the root as well. I also occasionally get an error that reads:

"An error occurred accessing your Windows SharePoint Services files. Authors - if authoring against a Web server, please contact the Webmaster for this server's Web site. WebMasters - please see the server's application event log for more details."

Usually restarting will fix it for awhile, but not long. Any ideas on what I might be missing?

 

[Sheco]

What details did you find in the server's event log?

 

[SpectralForce]

The only thing I get in the httper1.log file is the following type errors:

#Software: Microsoft HTTP API 1.0
#Version: 1.0
#Date: 2006-05-16 19:25:10
#Fields: date time c-ip c-port s-ip s-port cs-version cs-method cs-uri sc-status s-siteid s-reason s-queuename
with a whole bunch of Timer_ConnectionIdle - errors from the IP addresses of the computer trying to connect.

In the W3SVC1 and W3SVC2 log file it's mostly just GET and POST logs. I didn't see anything unusual, but I'm new to the IIS world. Is there something in particular I should look for?

 

[SpectralForce]

One thing that seems strange is that there are items in the log with with a timestamp of 2006-05-22 19:22:21 even though it's only 14:39 on the clock on the server. Am I missing something as to why the log lists the item 5 hours ahead?

 

[SpectralForce]

One other little fact that I just noticed is that if the person trying to edit disconnects and leaves it for awhile (I checked after 30 minutes) and connects again, it lets them connect. Could it be something to do with a slow network? Or a timeout setting somewhere?

Basically she saved the page, logged out and tried to login again and it gives the timed out error. But when we waited 30 minutes or so, then it started working again.

Thanks in advance for any ideas.

 

[SpectralForce]

Another thing I noticed is that if I use a local login to edit the pages (and not a domain user login) it doesn't seem to have the problem. Could it be something with the connection to the domain controller that is choking?

 

[safaritek]

The timestamp is likely using GMT, which is default unless you check the box to use the local time of the server.

In general, timeouts are due to network bottlenecks. It could be at the server or at the user's end. Is the server on the same network as the user?

 

[SpectralForce]

It is on the same network, but or network architecture is rather outdated and their is a problem with slow logins to the domain and other such issues. We are going to upgrade our backbone to 1000mb, so maybe that will help. I'll check out the GMT setting you mentioned and see if that helps. Thanks for the input. I was kind of out of ideas that would be quick fixes.