Splunk DB Connect 2 in a cluster

[stepheneardley]

I've a search head cluster running Splunk 6.4.1 and DB Connect v2.1.2. I've used this version successfully on standalone search heads but I've been having some problems with this cluster.

I installed DB Connect on my deployer as per the installation instructions, configured it then copied it down to shcluster/apps for deploying to my sh's.

I've noticed a few "weirdisms" since deploying...

1. When I open the app in my browser and navigate through different parts of the app it will sometimes refresh uncontrollably (seeming stuck in some kind of redirect loop) for a few seconds/minutes before the page loads completely and comes to a halt. A real pain but it doesn't stop me working.

2. The RPC server seems to constantly go up and down with the following lines in dbx2.log (Lots of lines...);

   `07/26/2016 14:00:13 [CRITICAL] [rpcstart.py] RPC server has been terminated abnormally with error [org.eclipse.jetty.server.Server@16267862 - STARTED].
   07/26/2016 14:00:13 [INFO] [rpcstart.py] proc.pid = 31324

07/26/2016 14:00:23 [CRITICAL] [rpcstart.py] RPC server has been terminated abnormally with error [org.eclipse.jetty.server.Server@16267862 - STARTED].
07/26/2016 14:00:23 [INFO] [rpcstart.py] proc.pid = 31378`

1. Scheduled DB inputs don't seem to be working i.e. I'm not seeing anything in my index and I can't find anything in Splunk logs to tell me what's happening.

I've considered upgrading to the latest version of DB Connect 2 but before I do can anyone confirm if upgrading will wipe my existing config please? I've got some extra clustering options enabled which are super useful such as sharing identity.dat across the cluster and replicating identities.conf and db-connections.conf between sh's.

Sorry for the long winded question(s). Is anyone else experiencing these kinds of issues with DB Connect 2 or is it just me?

[gendalia · Accepted solution]

Splunk DB Connect 2.1.3 is not compatible with search head clustering, I am unsure of the status of 2.1.2, but I believe I saw the same thing as (1) and (2) prior to my upgrade to 2.2.0. Version 2.2.0 is working for me with search head clustering.

I have never lost configuration across DB connect v2 upgrades, but I cannot speak to any particular set up.

[stepheneardley]

Thanks for this. Bit the bullet and upgraded and things are looking much better 🙂