How to prepare SCSP agent on Imaging computers? For example, we want to image a computer as our base image then SCSP agent is incuded on the image. Is there a conflict or issue when the SCSP agent is from imaged? No duplicate of computers registering on SCSP console?
Yes, it is safe to disconnect from the network, or you can use a lab network/subnet that does not have access to the network that the CSP manager is on.
CSP does not use a hardware ID for the nuts and bolts of the system. You can change an agent computer name as many times as you want, and CSP will not care, because it uses that GUID to represent the agent in the database. That is why when you change an asset's machine name, you will still see all the previous events from the asset in the Manager.
Theoretically, you can have 500 machines with the same name, and as long as they check in for the first time individually they will all get different GUIDs. The GUID is stored on the Asset, and if there is no GUID (or a "sisipsconfig -forcereg" is run), then the manager will give the asset (no matter what its name is) a new globally unique ID (GUID).
In a nutshell, do not allow the CSP agent to communicate with the manager when creating the Gold image, and you will be all set. Or, if you accidentally do, clear the GUID by running "sisipsconfig -forcereg" while the Gold Image machine is off the network.
Did you need any more assistance with this? If so, let us know. If not, please mark the issue as solved.
It is simple. Do NOT allow the SCSP agent to communicate with the Management Server when creating the Gold Image.
Here is how CSP registration works:
1) The agent is installed on the asset
2) Once the CSP service starts for the first time, it attempts to connect to the SCSP Manager using the preconfigured IP address/FQDN that you gave it when you installed. It also uses the .ssl certificate that you pointed to.
3) When the agent contacts the manager for the first time, the Manager gives the Agent a Globally Unique Identifier (GUID) that is saved on the agent locally, and entered into the SCSP Manager database.
That GUID is the key to each agent's individuality. If you allow an agent to connect to the manager during the creation of a Gold Image that you will use later in deployment, then you will essentially give out the same GUID every time you deploy that Gold Image, causing all the CSP assets to show up as the same machine in the console.
4) You can clear an agent's GUID by forcing a re-registration. The syntax is "sisipsconfig -forcereg". Note that Permission must be given to allow the use of the SCSP tools to do this.
So, as mentioned, the key to making a Gold Image is to not allow communication to the Management Server. You can accomplish this by either blocking access to the Management Server at the network level (disconnect the Ethernet cable, use a firewall, configure a switch, etc), you can give the install a dummy IP address which you can change later on (sisipsconfig -host xxx.xxx.xxx.xxx), or you can use an incorrect .ssl certificate, which would prevent the agent from communicating with the manager (then you can replace later using sisipsconfig -certfile).
Thank you Chuck for your response.
So, it is safe that I will disconnect from network while creating the Gold Image? How about the Hardware ID of the computer? Doesn't have conflicts if it has the same Hardware IDs?
Im going to test it on our test lab to test with these.
Again, thank you very much for your usual support.