Hi Chris and thanks for the post.

So in the first part of your post you say that the servers are showing up classified correctly as servers.  What protocol are you using in this case? SNMP? (seems you indicate that later).

Have you reviewed the steps in the First Time Setup tutorial?  This walks you through setting up OMSA on the servers and setting up SNMP on the target boxes.

Is it possible that these steps were done for the servers that are showing up correctly, but not done for others?

The subnets thing seems like it would be unrelated.  It usually goes back to the target...

Here are a few more things I check when getting "Unknowns":

Here are the important items to check when you are chasing down OME Discovery and Inventory issues.

1. Check FAQ 3.6 if inventory is not showing up.
2. Be sure OMSA is on the managed node (target)  if you are using SNMP/WMI (#3.3)
3. Be sure you have “accept packets” checked on your managed node (#3.4)
4. If discovering a 2008 server, see item #3.5
5. Be sure you check the community string, remember they are *case sensitive* (#3.4).  Did you type it in correctly in the discovery wizard?  Did you type it in correctly in the managed node SNMP properties?  Did you type it in correctly in the TS test screen?
6. Run the SNMP troubleshooting test for SNMP or WMI.  For SNMP the results *must* include the version of OMSA in the table.  For WMI the results *must* contain the words “Dell Server Agent” for the namespace returned.

Okay

Bullet point 2. OMSA is insalled and we are using SNMP and WMI.

Bullet point 3 & 5. Accept SNMP packets from these hosts (Dell OME server is listed). This is handled on all of our servers through Group Policy as well as the community string.

Also all DSM services are running.

Will start running Dell troubleshooting tools today.

Thanks for the updates.

Jus to confirm, you are saying that the configuration on all the managed nodes is identical and you get the exact same results from troubleshooting tool when you point it at any of the managed nodes.

Network latencies may play a role in discovery, Have you seen any ping timeouts for any of the subnets? You can try increasing ICMP timeout and retry values in the discovery wizard. You can also try reducing the discovery speed slider located under "Discovery Schedule". (It can be accessed by going to Manage->Discovery and Inventory and clicking the Discovery Schedule link on the left hand pane.)

It would be good if you can try one subnet first and then add few at a time.

I ran 4 test: Name resolution, Port, SNMP and WMI.

Name resolution passed by showing the FQDN of the server.

Port Number (1311) connected successfully.

WMI shows Present for Windows NameSpace, Dell Server Agent SnameSpace, PMI NameSpace and Virtualization.

SNMP shows the servername for MIB-II (System name), Server Administrator Version shows 6.5.0 and Inventory Collector Agent Version shows 6.5.0 as well.

Thanks for the updates. Keep us posted on the results from Troubleshooting Tool.

Yes the SNMP configuration is the same b/c it is pushed out through group policy. I have only pointed this one server for testing. I dont think SNMP timeout is the issue, b/c we do this everyday during off peak hours when there is almost no traffic on the T1 circuits at the sites where these servers are. The discover slider speed was set to halfway when these are failing, I can lower that more if you think it would help.

Thanks for the updates.

Lowering the slider speed will help if you are encountering timeouts. Especially ICMP timeouts. If you are not seeing those, you can keep slider speed as default. Do you know how many devices are there in the 80 subnets? There is a recommended hardware table in the online help under installing OpenManage Essentials section. You can check that to see if the OMe server meets the criterion.

As I mentioned earlier, discovering one subnet would be a good idea. That will tell you if you are able to discover all the servers in that subnet successfully.

Another thing to look at would be using SQL non-express edition if you have one. SQL express limits the memory to 1 GB and CPU to 1 socket and might cause performance issues.

Are you saying simply run a manual discovery process on that one subnet, or actually delete all Discovery Include ranges and plug them in one at a time?

It would be good to have one subnet under a discovery range to make sure all the servers in that range are discovered appropriately. If you have a separate range defined for a subnet, you can just kick off discovery or inventory for that range.