Best practice to install/configure DB connect in a distributed environment

[chang1800]

How should we install DB Connect in a distributed environment? We've read http://docs.splunk.com/Documentation/DBX/latest/DeployDBX/Otherdeploymentconsiderations but still aren't sure.

We have 2 search heads with search head pooling.

We have 3 indexers that are all receivers and search peers of the search heads.

We have 1 utility box we can dedicate to DB Connect.

We're not sure if it should be a search head (not pooled) or a heavy forwarder.

We think we'd like to make the utility box a heavy forwarder dedicated to DB Connect, but are there any downsides?
Can we then search the db connect data from our pooled search heads?

Will the db connect split across our 3 indexers be fully searchable?

Do we need to install anything on the search heads?

Also, we have 1 deployment server. From the docs, it looks like we would not use that to manage any DB Connect config. Is that right?

[banchio]

Another way of approaching this problem could be the following:

1) If you plan to use DBConnect for data enrichments, only (Db lookups and so on, mainly search-time) you could install it on a Search Head.
2) if you plan to rely strongly on database inputs, you should evaluate expected volumes:
Small => install DB Connect on a search head and make it forward events to the indexer(s)
Medium => install DB Connect on a dedicated Splunk heavy forwarder

Large => install DB Connect on multiple Splunk heavy forwarder

source: http://docs.splunk.com/Documentation/DBX/latest/DeployDBX/Otherdeploymentconsiderations

[ziegfried · Accepted solution]

In order to use database inputs, the recommendation is to install DB Connect on a (dedicated) heavy forwarder. This forwarder should be configured to forward the data to all indexers using load-balancing.

In order to use reporting features, such as lookups, ad-hoc queries or dboutput, DB Connect needs to be install on the search heads as well. Setup on search heads with pooling enabled requires some additional steps, see: http://splunk-base.splunk.com/answers/68598/java-access-bridge-not-running-with-search-head-pooling/...

[hemendralodhi]

How can we use reporting feature as we are using search head clustering and as per doc it is not supported there?

[chang1800]

ziegfried, Thanks for your speedy answer.

[pkapila]

Hi - Can you point me to install steps for DBConnect on Windows heavy forwarder?