we have an application that insists on writing relevant log information to an MS SQL DB instead of decent files.
This of course means I need DB Connect to read the logs. I am not interested in direct lookup from the search head.
So how do I set up the Heavy Forwarders in our environment?
I have two Indexers, a search head and a deployment server.
I definitely do not want additional hardware or VMs.
My plan is to install the HF and DB Connect on the application server just like I would install a UF.
And then have it connect to the applications DB System via JDBC. The Heavy forwarder should of course also forward the local logs from the application server and not just the log extracted from the DB. The application server has some extra capacity to handle the load.
I would prefer to manage the HF just like a UF via deployed configurations without any GUI active on the HF and no local configuration apart from the initial install of HF and DB connect and the setup of the deployment client configuration.
Is this possible or do I absolutely need the GUI on the HF to set up the DB Connect or other things?
Can a configuration for the DB connect be pushed via some other deployment app (as the DB Connect itself is not deployable)?
Ok, tried it and failed.
Key issue for not being able to use the deployment server is the DB password. The pw data sent by the deployment server needs to be hashed with the seed from the UF. Yes, that could be hacked but is not very practical.
And it seems that when using rising column mode some file gets initialized via the GUI which I have not been able to create manually.
And currently the GUI is broken thanks to Python 3 issues ;-(
Your plan is ok. You could do is just as described above.
Anyhow personally I propose to use GUI for configuring inputs and connections. It also contains Health monitoring part. Of course You could/should install that also to your SH for Monitoring and see current status of DBConnection. There haven't been any reason to use those lookups or other additional features in SH. Don't use those inputs on SH side, those must do on HF side on distributed environment.