SmartStore S3 data replication

goldeneye1117 — Wed, 29 Nov 2023 19:13:03 GMT

I have been testing out SmartStore in a test environment. I can not find the setting to control how quickly data ingested into splunk can be replicated to my S3 bucket. What I want is for any data ingested to be replicated to my s3 bucket as quickly as possible, I am looking for the closest to 0 minutes of data loss. Data only seems to replicate when the Splunk server is restarted. I have tested this by setting up another splunk server with the same s3 bucket as my original, and it seems to have only picked up older data when searching.

max_cache_size

only controls the size of the local cache which I'm not after

hotlist_recency_secs

controls how long before hot data could be deleted from cache, not how long before it is replicated to s3

frozenTimePeriodInSecs, maxGlobalDataSizeMB, maxGlobalRawDataSizeMB

controls freezing behavior which is not what I'm looking for.

What setting do I need to configure? Am I missing something within conf files in Splunk or permissions to set in AWS for S3?

Thank you for the help in advance!

Re: SmartStore S3 data replication

richgalloway — Wed, 29 Nov 2023 20:04:14 GMT

Data is written to SmartStore (S2) as soon as it rolls to warm. On a test system, it's typical for hot buckets to not roll to warm until the indexers restart. On a production system, however, that should happen at least once a day. Hot buckets are never written to S2. There is no setting to give you instant replication to S2. In an indexer cluster, hot buckets are replicated to other indexers almost immediately.

topic SmartStore S3 data replication in Knowledge Management

SmartStore S3 data replication

Re: SmartStore S3 data replication