I am using the kibana puppet module, and having some issues with plugin installation. I am on an isolated network and unable to install any plugins except from local file; however, not from the puppet manifest. This is the manifest that I am using:
When I run the manifest, I get the following error:
Error: /Stage/[main]/uss_install_elk/Kibana_plugin[searchguard]: Could not evaluate:: Attempting to transfer from file:///path/to/zip
Transfering 1845207 bytes.................
Transfer complete
Retrieving metadata from plugin archive
Extracting plugin archive
Extraction complete
Optimizing and caching browser bundles...
@tbmorris78 are you copying the files to the server before attempting to use the kibana_plugin module with the "file:///" syntax? Also, which version of Kibana are you using and which OS are you provisioning?
Sorry about leaving that out. I am using Kibana 5.2.2 on CentOS 7.3. Yes, before executing the kibana_plugin resource, I have a file resource that copies the file over. It would be nice if I could point it at a puppet:// resource, but that is for another venue.
As I mentioned in some messages between us, this is likely a timeout issue, but that's somewhat more tricky to debug, and I think there's a way to get around the error as a workaround since I'm not sure at the moment how to workaround waiting for the long command execution time.
The kibana_plugin resource type checks to see whether a plugin is installed by looking for a matching plugin name in the plugins directory. If the plugin ends up at /usr/share/kibana/plugins/searchguard, then the resource should be kibana_plugin { 'searchguard': in the example manifest you provided.
The example manifest and log output looks different to me, as the resource is named searchguard-kibana but the logs say it is searchguard - it should be searchguard in the manifest, and then hopefully Puppet will be able to detect that it's installed, even if the command exceeded some Puppet timeout and completed in the background.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.