I am trying to 'upgrade' Splunk from 7.2.5 to 8.0.3. Splunk is running on a RHEL7 VM in a docker container from Splunk. (We not actually upgrading Splunk, we are moving to a new container on a new VM.) Through automation, we had modified our container's etc/system/local/inputs.conf to run with SSL according to the Splunk documentation, and in 7.2.5 this works.
In 8.0.3, we are finding that configuration entries in inputs.conf are being erased whenever we restart docker. (/opt/splunk is a folder mounted in the container so that it persists.) Splunk is not 'restoring' the file (for example, from the ../defaults folder) - from testing, we've discovered that some comments do survive, but the configuration entries for SSL are being deleted and Splunk 8 is not running using SSL.
server.conf is also getting clobbered.
Anyone else notice this behavior?
Before restart:
[default]
host = edb999320984
# BEGIN ANSIBLE MANAGED BLOCK
[splunktcp-ssl:9997]
disabled = 0
[SSL]
serverCert = /opt/splunk/etc/...
requireClientCert=false
# END ANSIBLE MANAGED BLOCK
After restart, all that remains is:
[splunktcp://9997]
disabled = 0
# BEGIN ANSIBLE MANAGED BLOCK
One other thing we notice is that with Splunk 7, the files are owned by 999:999. In Splunk 8, the owner/group is 41812:41812. However, adjusting for that, our config changes are still getting clobbered.