No, this will not create the index on the indexer, only on your search head.
What I did @one costumer, was to create a script that could be called with an argument like
index=foo and this scripted looped through all
splunk_server and search heads and created over REST API calls the index on all Splunk servers.
hope this helps ...
Thanks that make sense , so basically you need to create new index on all indexers?
Also does it make sense to use Main index for everything - i was thinking to split the event streams in various indexes right from start - any suggestions?
I usually create new indexes for new 'data areas' - and then split it further by sourcetype.
so for instance for systems metrics i'd create index system, and then tag linux sourced data with sourcetype=linux, windows would be sourcetype=windows etc.
for app data i'd create separate index (ie. index=appdata) etc etc.
3 reasons for it:
1. don't always need to search through all the data
2. data access permissions are set on per index basis
3. data retention times are set on per index basis.
hope that helps.
That totally makes sense - qq, when you index data based on source type events will go to individual data streams and not main - but lets say i have a user who want to see events from Server XYZ and he is L1 person and dont know much on indexing - from his point of view when he will search on search panel with device XYZ he will not get any data back since default index is main on search right?
Or do i need to assign roles based on what index they can search on ? i dont expect him to write a query like Hostname=XYZ index="App"
Creating different roles for different type of access is recommended. You can then set the default index to be searched for each role.