Closed VatsalJagani closed 1 year ago
Currently following are the problems for distributing limits.conf and db file to Indexer Cluster:
If you make the following changes you can avoid using limits.conf, it will replicate to indexers and works in cloud. bin/mmdb_utils.py 22 MMDB_FILE_NAME = 'GeoIP2-City.mmdb' 25 DB_DIR_PATH = '/opt/splunk/etc/apps/search/lookups’ 117 #self.update_mmdb_location()
@sclapper - I don't think on indexer the location of the lookup would be same. It would be under /opt/splunk/var/run/
/opt/splunk/etc/apps/search/lookups is just the path where you put GeoIP2-City.mmdb, just need to write the mmdb to ..search/lookups/GeoIP2-City.mmdb and it will replicate.
@sclapper - Yes that is accurate it will replicate, but the location on the indexer would not be /opt/splunk/etc/apps/search/lookups. It would something like this: /opt/splunk/var/run/searchpeers/splunkui111-1324253269/apps/search/lookups. Path would have dynamic number with it (replicated bundle number)
So how would you assign a path in the limits.conf for the indexer? Also, the path for the indexer and search head would be different so you need to write different limits.conf for search head and indexer, which is not possible in Splunk cloud.
This method doesn’t require using limits.conf, Test it out.
When you update the database using this method (as a lookup) https://docs.splunk.com/Documentation/Splunk/9.0.1/SearchReference/Iplocation#Updating_the_IP_geolocation_database_file
It gets written to the splunk_dir/etc/apps/search/lookups dir and it gets distributed to the indexers. I believe this is a new feature in Splunkv9 but has been available in Splunk cloud for a while longer.
@sclapper, @kyndsafian - Thanks for your suggestion and comments. We're working on this to improve for both enterprise and cloud by using this new feature.
@sclapper , @kyndsafian - Thanks for your suggestion and guidance.
We have implemented the change here - https://github.com/CrossRealms/Splunk-App-Auto-Update-MaxMind-Database/pull/13 We have tested the approach it's working in SHC.
Currently, we are resolving one other issue with SHC - Other SHs opening the setup page even after App setup is completed on one SH, and app.conf is_configured parameter is also replicated to other SHs. Once that is resolved, we'll create a new release of the App.
Really appreciate your comments here.
App version 2.0.0 has been released with proper support of the Search Head Cluster, Distributed environment, and Splunk Cloud Classic and Victoria experience. https://splunkbase.splunk.com/app/5482
Thanks for everyone's suggestions and support. @sclapper , @kyndsafian