In ArcGIS Server, you can create and publish web applications with no problem, but the map itself does not appear when you view the web application. Instead, you see all the ArcGIS map tools, and where the image should be you see a broken image X. ArcMap and ArcCatalog can connect and view the map images with no problem, so the problem lies with the web server, IIS.

Running Windows Server 2008 64bit, IIS7 with ArcGIS Server 9.3. The fix was the following had to do with the ASP.NET mode the application pool is configured to run, specifically the managed pipeline mode. You can choose integration or classic mode. By default, IIS7 uses an integration mode. If you change the managed pipeline mode to classic, then the application maps appear as expected. To change the mode, go into IIS Manager, click on application pools, and view the settings to see the managed pipeline mode.

What is this so called ASP.NET managed pipeline mode? IIS6 works in a way that when a request is made to ASP.NET, it first goes to IIS, and then forwarded to the ISAPI DLL so that ASP.NET can process the request. In IIS7, the request for ASP.NET is sent and both IIS and ASP.NET have the ability to process the request.

I’ve only seen this problem once. I’ve had servers run in integration mode with no problem. So it could be an issue with creating a server off of an image and running sysprep, and DCOM doesn’t correctly register.

In ArcGIS Server manager, you are unable to create a web app. You receive the error “An error has occured on the server. For details, please check the event (application) log on the web server. Unable to retrieve ArcGIS install directory from the registry.

The reason is because 32-bit IIS/ASP.NET is not enabled. We were running IIS7 on Windows Server 2008 64bit. Once you enabled 32-bit IIS/ASP.NET, everything works as expected.

When installing ArcGIS 9.3 on Windows Server 2008, you have to  make sure UAC is disabled. If not, the install will go fine, but when you get to the ArcGIS Server Manager to change the user store to SQL or Windows Users and Groups, it will show users and groups grayed out, and it will not let you change it. Once UAC is diabled, it will default to windows users and groups.

YES.

Currently ESRI is only supporting ArcGIS Server 9.3 and Apache for reverse proxy. Over the past few days, I was able to use Microsoft ISA Server 2006 Enterprise Edition as a reverse proxy to allow external users to access the ArcGIS Server. The best part, we used forms based authentication and Active Directory. The external customer comes in, autheticates to the ISA Server using AD credentials, and if successful those credentials get passed on to ArcGIS Server. The communication side works like this. User comes in on port 443 SSL, ISA server communicates with the ArcGIS server to a user specified port, we setup 443, but it will work with say 444, for more security you change that port. The reverse proxy solution is the most secure way to allow users from the external side to access the internal resource.

If you do troubleshooting for remote users, one that i’ve come across as a problem is asking the user to unscrew their laptop casing to find out how much RAM is in one slot. And if you have a DELL, most likely you have one DIMM slot on the bottom of the laptop, and then you have to open the top, remove the keyboad, to find the 2nd hidden slot.

Have no fear, I just ran into a piece of software, CPU-Z by CPUID that does EXACTLY that. It reports detailed information about the slots, how much RAM, etc. It also gives very good information regarding the CPU. Check the software out here http://www.cpuid.com/cpuz.php.

Here are some key notes to remember when you want to migrate your virtual server images from Windows Server 2003 running Virtual Server 2005 R2 to the Windows Server 2008 Hyper-V.

• Hyper-V uses it’s own new tools called integration services. Therfore, before moving the virtual server image to the new server, UNINSTALL the virtual server additions.
• The network adapters will most likely change when moving, and if have static IPs, you will get a conflict when you try to assign the same IP on the Hyper-V. Best bet is before moving the VHD to the new server, change your network settings to default
• In virtual server 2005 you could start your virtual server on the SCSI adapter. In Hyper-V this is not supported, the main boot device has to be IDE. So either create virtual server image with the system OS drive as the IDE and misc apps on the SCSI device, or just live with IDE.

The DELL ProCurve 2724 is a decent web managed switch. What I was trying to accomplish was setting up VLANs on my main switch and passing the information to the 2725 through a TRUNK.

A trunk is used so that instead of taking a network cable from one switch’s VLAN and running it to another switch (vlan 1 to vlan 1) you can use one port to hold all the VLAN information. If you didn’t use trunking, then each VLAN that you wanted to communication between the switches would EACH need a cable running to the VLAN.

VLAN1 is the default VLAN. You cannot receive TAGGED packets to VLAN1. Other VLANs you can receive tagged packets. I wanted to setup 3 VLANs.

VLAN 1 - Ports 1-8
VLAN 2 - Ports 9-17
VLAN 3 - Ports 18-24

First you go into your Dell management web gui and click on VLAN membership. Here you create VLAN 2 and VLAN 3. Next you UNTAG packets on the VLANS for the specific ports, so VLAN1 would have  U by ports 1-8. I will be using PORT 1 as the trunk, so for VLAN2 and VLAN3, make sure you specify port 1 as the trunk, and it will show a T.  VLAN1 cannot cannot receive tagged packets, so you cannot make port1 on VLAN1 the T, it will remain U.

The switch that will be connected to the Dell Powerconnect 2724 is a HP ProCurve. The trunk port on this end is configured to send UNTAGGED packets to VLAN1, and TAGGED packets to VLAN2 and VLAN3. This is very important, or nothing will work.

Now this line of Dell switches are a little different. You created the VLAN you think and specified which ports are associated with which VLAN under the VLAN membership. For some reason, we are NOT done. Click on the VLAN port settings. Here you will need to AGAIN specify which VLAN each port is associated with, by entering in the VLAN # in the PVID.

Sometimes you will  have a username for your ftp as user@domain.com. Now the problem with embedding this URL in a web site or email is that there are 2 @ symbols. For example, ftp://user@domain.com:password@ftp.domain.com. the fix? use the %40 which is the @ symbol replacement. So ftp://user%40domain.com:password@ftp.domain.com

This has been going on for months. January to be exact. A server running Windows Server Datacenter 2003 x64 edition, A Dell PowerEdge 2950. Dell has NO clue why this is happening. i’ve called them probably on an average of once a week and each time they say it’s been escalated. Anyone else having this issue?

Update 7\31\2008

I got so frustrated with the issue and there was no fix after 7 months. They sent me out a NEW server. A Dell PowerEdge 2970 with the AMD Opteron. I decided to install Windows Server 2008 64bit edition on here.

The did narrow down that the problems they were having were the Generation III servers of the Poweredge 2950 and 1950 product line.

When you use the command w32tm /resync, you receive this error. Why? A number of reasons. You do need to open port 123 UDP on your firewall for the domain controller so it can reach the external time source. If you want to sync your clients to your DC, you probably edited a domain policy to allow this to happen. The catch what I noticed is that it’s most likely that your DC is part of this domain policy. By having a NTP client configured there, the DC will not sync to the external time source. Mistake, yes. Fix, well, create another policy JUST for your clients and specify your NTP client information there conneting to your DC. As for the domain policy, put all time related configurations as NOT CONFIGURED. By doing this, I was able to issue the w32tm /resync command succesfully.