Hi everyone,
I have 3 indexers (in a cluster) located on Site A. The current replication factor (RF) is set to 3.
I need to move operations to Site B. However, for technical reasons, I cannot physically move the existing data — I will simply have 3 new indexers at Site B.
Here’s the plan I’m considering:
Launch 1 new indexer at Site B.
Add the new indexer to the existing cluster.
Increase the RF to 4 (so that all raw data is fully replicated across the available indexers).
Shut down all 3 indexers at Site A.
Decrease the RF back to 3. (I understand there is a risk of some data loss.)
Add 2 additional new indexers to the cluster at Site B.
My main concern is step 5 — decreasing the RF — which I know is not best practice, but given my situation, I don't have many options.
Has anyone encountered a similar situation before?
I'd appreciate any advice, lessons learned, or other options I might not be aware of.
Thanks in advance!
If you really dont want to go down the multisite route then you could keep your RF at 3 and slowly introduce new indexers in the new location, offlining one from the old site as each new one is added, although I really would recommend the multisite approach personally...
Here is what you would do:
Why This Works as an approach
🌟 Did this answer help you? If so, please consider:
Your feedback encourages the volunteers in this community to continue contributing
If you really dont want to go down the multisite route then you could keep your RF at 3 and slowly introduce new indexers in the new location, offlining one from the old site as each new one is added, although I really would recommend the multisite approach personally...
Here is what you would do:
Why This Works as an approach
🌟 Did this answer help you? If so, please consider:
Your feedback encourages the volunteers in this community to continue contributing
Hi @m_zandinia
I dont think upping your RF to 4 is the right approach here, you should probably treat this as a single-to-multisite cluster migration, even if you are going to deprecate the old site afterwards.
There are some useful docs at https://docs.splunk.com/Documentation/Splunk/9.4.1/Indexer/Migratetomultisite covering thism,
Directly decreasing the Replication Factor (RF) when indexers holding necessary copies are offline can lead to data loss because the cluster manager may still believe those hosts exist.
Migrating data between sites using multisite replication takes time and network bandwidth. Monitor the cluster status closely (`splunk show cluster-status --verbose` or the Monitoring Console) to ensure replication completes before decommissioning the old site.
Plan your site_replication_factor and site_search_factor carefully based on your desired redundancy and search availability during and after the migration.
Useful Documentation Links:
🌟 Did this answer help you? If so, please consider:
Your feedback encourages the volunteers in this community to continue contributing