Slow Windchill browser and Creo losing server connection

[mspencer]

Our Windchill 10.0 M030 install is having issues today...the web browser interface is 'working' but very slow and intermittent. Users interfacing through Creo are continually losing connection to the server to the point where it is impractical to work. Simply connecting to the server thru Creo causes a severe hang (5 minutes), and displaying the workspace afterwards is very lagging.

Our Windchill server machine and SQL machine are not resource limited (cpu/memory/hdd/network), and are generally idle while the sytem appears to hang. Nothing changed with our general system setup (WC 10.0 M030 / SQL / Creo 1.0 M030) for over a month. We did have an unexpected restart of our SQL machine late last week due to an unintentional windows update.

There are no obvious log errors on the method server or manager logs, nor the apache logs. We tried clearing the tomcat cache and restarting both the SQL and Winchill server machines, with no improvements. The server status page shows the server manager and method server as up and responding with no big flags. Attached is a screenshot.

Our VAR is looking into the issue, but hoping we could get some suggestions on how to narrow down the problem, or potential causes.

Thanks for your help!

[dineshmech]

Switch to Mozilla browser and check if that helps for users accessing only via web. For Pro/E users, enable embedded browser as Mozilla via config.pro option windows_browser_type=mozilla_based_browser. I hope this helps...

[TomU]

Did this already. No change. The issue is not with standalone browsers since these are not looking at local workspace content (only server side workspace content). This behavior only happens when using the embedded browser. Performance with the embedded browser does not change regardless of which browser is used (using the config pro option "windows_browser_type").

Tom U.

[TomU]

Sorry, wrong list. Ignore my response to this thread please. I thought we were talking about something else.

Tom U.

[sscole1]

Make sure you no longer have a "dm_network_threads" config.pro option set (depricated in later releases of Creo) and you may want to try setting "

CS46526

[jellis]

Max....we set up WC10 m030 in early May. Ours is running in Oracle. Similar problems arose with performance. It was very sluggish and non-responsive. Through many hours of turmoil we finally ran across a Microsoft article that suggested installing a service pack on the server os. We installed the service pack andWindchill 10performance is quite acceptable.

[mspencer]

Update to include the resolution, we found two items which fixed the issues.

First, our IT consultant went on a virus scan install spree one weekend and smacked AV all over our WC and SQL servers. We promptly disabled this.

Second, which took a bit more asking and trying is setting the 'KeepAliveTimeout' parameter in the Apache conf file. Yes this is addressed in several PTC support articles...but somehow our VAR didn't set the parameter, nor identify this as the problem. It is now set at 30, and our problems have all but vanished. Kudos to offical PTC tech support for quickly identifying and correcting the problem!

[jessh]

You may have to set this as high as 61.

The reason for this is a serious and unresolved bug in Internet
Explorer's handling of HTTP keepalives. Use of any other browser will
also resolve the issue.