I have a deployment server and I have created a serverclass to send apps to my clustermaster. I want to receive the apps in master-apps instead of apps folder.
I have used this on my clustermaster:
[deployment-client]
repositoryLocation = $SPLUNK_HOME/etc/master-apps
serverRepositoryLocationPolicy = rejectAlways
I'm getting the apps but every minute the clustermaster tries to get a fresh copy from the deployment server, it is like if the checksum was being validated against apps, not master-apps.
If I use the regular apps repositoryLocation everything works fine.
Do I need to modify repositoryLocation in any other place:
this is what I'm getting from serverclass.xml on the client:
``
any thoughts?
I have seen clients do this by setting up a cron job to run every hour/half-hour:
/opt/splunk/bin/splunk apply cluster-bundle
If there have been no changes to the folder's contents, applying the cluster bundle doesn't do anything (Since the hash of the folder's contents match the hash of the currently applied cluster bundle, so there is nothing to and the bundle-apply process exits).
There is the risk that with automating like this that if the deployment server sends down certain types of changes down to the cluster mater, it'll initiate a rolling-restart. This link will detail what might cause that: Restart or reload after configuration bundle changes?
I'd recommend testing this before going live with the change as there may be something specific to your environment that may cause issues.
I downvoted this post because this does not answer the question of why the system is checking in every minute and updating the files.
Hi .. Im having similar issues .. did you end up resolving this problem? If so, how?
Thanks
Shaun
Hi @shaun.dyble
If you're experiencing the same issue covered in a question, be sure to click the "ME TOO" button just below the question title on the left side. This will show multiple people are experiencing this and help improve the visibility of the post.