If you are using Altiris Out-of-Band Management v6 with Intel SCS 3.x (check the AMTconfig windows service) - and you keep getting a series of w3wp errors on your Altiris console.... this article might be of interest to you.
Of course - try as I might to repeat this exact error in the lab, or in talking with customers using the solution in production environments - it was tough to repeat the error. The exact circumstances that generate the error were not available, yet the occurrence does happen due to a memory leak in amtconfserver (i.e. Intel SCS). If you're really ambitious and examine the memory dump file - there will be references to mod_gsoap.dll?services or services.dll within the AMTSCS web directory.
Since the w3wp fault due to amtconfserver may also affect Altiris web application pools, the error can have dramatic effects within a production environment. The workaround is to isolate the web app pool for Intel SCS.
The challenge is that Intel SCS development is moving forward - the recommended path is to upgrade to version 5.x or 6.x. - however, the supported version in an Altiris 6.x environment is 3.x. So - to stay within compliance, this article provides one workaround approach.
The following image is from a customer production environment. The server name has intentionally been blurred to protect the innocent.
To isolate the web app pool, open IIS Manager (one approach is right click My Computer, Select Manage, navigate to the IIS Manager).
Right click on Application Pools under IIS Manager and select New > Application Pool
Provide a name for the application pool and assign the DefaultAppPool template - as shown below.
With the new application pool created, assigned the SCS web instance to the SCS application pool. This is done by expanding Web Site, right click on AMTSCS, select Properties, and adjust the Application Pool from DefaultAppPool to SCS (the newly created app pool)
Restart IIS to ensure the app pools are cleared, only the desired websites running per app pool, etc.
The SCS app pool can be adjusted to recycle memory or have some resiliency by increasing the number of web gardens. Per the following screen shots, start with 100 MB virtual memory and 40MB used memory. Also, set the web garden to a value between 6 and 12. Adjust settings while monitoring performance... until you reach a satisfactory balance.
To check the actual resource utilization of the w3wp process associated to this particular web app pool, try the Sysinternals Process Explorer (http://technet.microsoft.com/en-us/sysinternals/bb896653.aspx)
For each defined and in-use application pool, and individual w3wp instance will appear. Thus it may require a few attempts to determine which one is related to the SCS app pool defined above. The following screenshot shows the app_pool_id in the environment tab is associated to SCS... and the other two w3wp instance refer to the DefaultAppPool and AeXHD
With the web garden value set to value above 1 - the SCS app pool can more gracefully fail without impacting the overall webservice. In addition, with the AMTSCS web instance running in a different application pool from the Altiris web instance - w3wp faults will be isolated. Note that for each web garden value of the SCS app pool - an associated w3wp instance will be started\running.
As noted above - the underlying issue is a bug in Intel SCS - which will appear as AMTconfserver in the event viewer along with w3wp fault, the AMTconfig windows process, etc. The bug is a memory leak in SCS 3.x, which is fixed in SCS 5.x and above. However, the newer versions of SCS are not tested\validated with Altiris 6 environments... and development\support on SCS 3.x has ceased. Thus the workaround.
Hope that helps. Interested to hear if others are experiencing this error... since I wasn't able to repeat it, yet did receive enough confirmation that it existed to provide this article.
The opinions expressed on this site are mine alone and do not necessarily reflect the opinions or strategies of Intel Corporation or its worldwide subsidiaries