How often report accelaration check the new data set in HDFS ?

[melonman]

Hi

I am using Hunk and configured report acceleration saved searches. report acceleration automatically checks where there are new data sets that have not been cached.

My Question is how often the report acceleration feature checks new data set in HDFS, and its mechanism.
Is there default value of interval for report acceleration to check the new data sets?
How does the report acceleration keep track of cached/fresh data set?

Appreciated if anyone can point me to the information or readings about this.

Thank you.

[Ledion_Bitincka · Accepted solution]

It defaults to every 10 minutes, but it can be set on a per saved search basis from the conf file, for more info look at savedsearches.conf. Note that there is a limit to how many report acceleration summary generating searches can run concurrently (defaults to 50% of scheduler's capacity), which means that for busy systems the 10 minute period is only a rough approximation.

auto_summarize.cron_schedule = <cron-string>
* Cron schedule to be used to probe/generate the summaries for this search