We have a master cluster, 2 search heads and 2 indexers in our setup. We have done development on a standalone server and are now moving the configuration to the cluster. I put the dbx app out in the /opt/splunk/etc/master-apps directory so that it gets distributed to both indexers, but we are seeing some issues:
- Configuration bundle distribution fails from the UI on the master cluster server. It complains about invalid stanzas and spec files. If I do a distribution from the command line with the --skip-validation option, everything goes out and the db connections gather data like they are supposed to. Why is this complaining when performing validation?
- Unable to list the database inputs on the indexers. When I click on Settings->Data Inputs, I see that there are 2 Database INputs listed which is correct. When I click on the database inputs link to list them, I get "An error occurred completing this request: In handler 'dbx-monitors': Cannot call handler 'dbx-monitors' due to missing script 'resthandlerdbmon.py'". I verified that this file exists under the /opt/splunk/etc/slave-apps/dbx directory, so not sure why this is erroring out.
- Unable to list the external databases on the indexers. When I click on Settings->External Databases, I get "An error occurred completing this request: In handler 'databases': Cannot call handler 'databases' due to missing script 'resthandlerdbs.py'". I also verified that this file exists under slave-apps.
- Lastly, I was under the impression that if you had two peer servers in a cluster synchronizing the same indexes that there would be no duplicates. We are querying data from the same source db tables and putting them into the same indexes on both peers and we are seeing two of every row. Is there a setting we need to set to deduplicate?
Sorry, DB Connect is not currently certified or supported with clustering. I'll open an issue to see if we can increase our coverage for the next version of the product.
So, it is about indexer cluster, but what is about Search Head Cluster? Is it possible to use DB connect for lookups in external database in SH cluster?
Its not "supported" in shc mode but I have gotten it somewhat working.
There are a couple of ways to do this.
Either have a locally maintained install on each member OR use a shared splunk.secret file on each member and use a deployer based unique database.conf file. You will need one of the systems to originally generate the password hashes that are used so they are consistent across all the members.
I have run into a couple of weird issues I am trying to resolve with this configuration however.