Splunk 103 Flashcards
What is Indexer Clustering?
Clustering is where multiple indexers are connected in order to maintain multiple identical copies of data. Clusters featuer “automatic failover”, which simply means when or if one indexer fails, the others will pick up the slack and maintain continuity in its activites.
This means:
- Data is protected from sudden loss
- More copies are available for users who are actively searching
- The above acitvities will continue even when an indexer goes down
What determines the number of copies kept withint a cluster?
Replication Factor
What is the Replciation Factor?
This determines how many copies are maintained within an indexer cluster.
Default RF is 3.
What is the Search Factor?
This determines how many copies in the cluster are immediately searchable
Default SF is 2.
What is the minimum amount of indexers that you have to have in cluster?
3
What are components of a cluster?
Cluster Peer (Peer Node) Cluster Master (Master Node)
What are Cluster Peers?
Peers are the indexers that are in the cluster. They recieve and index incoming data, and replicate it to other peers. They respond to incoming searches by supplying search results.
What is Cluster Master?
It manages cluster activities (such as adding peers, distributing configurations, determining number of copies to maintaing)
It maintains memory of peers, their buckets and configs, and tells search heads where to request data
How does distributed search work in the cluster?
Search head “asks” Master Node in which indexers it should should search for the data it is searching for, and then it accesses those indexers.
What are benefits of clustering?
Data availability and fast recovery
Easier overall administration:
- Coordinated indexer configuration management
- Automatic distributed search set up
- Elastic indexer discovery
- Indexer health dashboard on Cluster Master
Scalability of Indexing
No additional license cost for data replication
Data fidelity
Data Resilency
Disaster Recovery
Search Affinity
What are cons of clustering?
Increased storage requirements
Increased processing load (depends of RF and SF)
Requires additional Splunk instances:
Minimum: RF + CM + SH = # of insances required
REcommended: More than RF, and multiple SHs
Indexers require the same OS and versions
Requires cluster specific deployment management
What are configuration bundles?
A set of configuration files and apps common to all peers.
Where do configuration bundles reside on cluster master and cluster peer
Cluster Master:
$SPLUNK_HOME/etc/master-apps
Cluster Peer:
$SPLUNK_HOME/etc/slave-apps
What are some of the configuration changes that require a restart?
Changes to indexes.conf, inputs,conf
Changes to a home path in indexes.conf
Dleeting an existing app
What are some of the configuration changes that do not need a restart?
Adding a new index or a new app with reloadable configs
Changes or additions to transforms.conf or props.conf