solman_setup: Managed system Host configuration - Enable Diagnostics Agents on-the-fly is greyed out

[#1]

As oftentimes said, a picture says more than 1000 words:

I tried searching on google and on the market place - it looks that we are the only ones in the universe facing this problem 😞

We tried:
- restarting the diagnostics agent
- saphostexec -restart

Nothing helps, the option just stays gray and thats it. Nothing... 😕 Please HELP!

[Vera_Jiang88]

Dear Customer, The on-the-fly feature should only be available for physical or virtual hosts. It appears that the type of host is currently unclear. Is there any problem with the Outside Discovery? Please refer to KBA 3232202 for more information on DatabaseOutsideDiscovery / ComputerSystem. Furthermore, is the SAP Host Agent status satisfactory as mentioned in KBA 1890035? Best regards, Vera Jiang

[mamartins]

When you install the SOLMAN Diagnostics Agent, on the very first SWPM forms, there is one asking the name of the server. It's mandatory that you leave it empty (it's pre filled, so you need to remove the hostname), otherwise the agent-on-the-fly functionality will never work!

[symon_braunbaer]

Thank you, mamartins,

I know this already. We are installing the Diagnostics Agents with automatic scripts, otherwise installing Diagnostics Agents would be the only thing we are doing 😄 We do not have a hostname in the .properties file. Furthermore, this problem is only on the Servers where the ASCS / ERS cluster is running. On the AS Servers we do not observe this problem and I could configure Agents on-the-fly without any problem. And the Diagnostics Agents were installed by the same automated manner.

For some reason, there are no FQDN entries in the solman_setup Host configuration for these 2 Servers:

On all other servers we see a ton of entries in the FQDN column. Yesterday I was checking the name resolution:

nslookup hostname
nslookup <resulting_IP>

There were some discrepancies, I have fixed them, but the situation is still the same. Do I have to restart the Diagnostics Agent ?
I think it should pick up any modifications automatically ? Otherwise I am really running out of ideas...

[mamartins]

The name resolution (DNS, hosts local file) is also critical to this to work. I had seen situations where multiple restarts of the DAA were necessary in order to operate properly.

[symon_braunbaer]

Our resolution order is first the local /etc/hosts file and then DNS. Would you recommend anything different ? But anyway, it is working on all other systems, just on this one it isn't

[symon_braunbaer]

We just finished a HUGE troubleshooting session with the OS guys. Still no solution !! :-x But I noticed one thing. Things are not looking good in LMDB for the 2 ASCS / ERS hosts:

I also checked in the SLD, this is fine, it seems to contain all data as other systems. But as we see in the above screenshot, LMDB is not fine. Please kindly suggest how to fix this. Many thanks!

[mamartins]

Confirm that you have only one ASCS instance (as a cluster resource, so can run it on any of the cluster nodes) and one ERS instance per cluster node (outside cluster, as a local host service and can share the same instance number).

Many times the LMDB have errors on these more complex landscapes, so manual data edit is necessary. The steps needed before manual updating:

• On the ABAP side, run t-code RZ70 and force an update to the SLD
• Next, enter on the SLD and check if the system data latest update is changed
• Then go to the SOLMAN, and execute the program RLMDB_SYNC_TECHNICAL_SYSTEM on the T-CODE SA37 to get the latest data from the SLD
• Enter the LMDB again and check. Enter manual data as needed.