I'm facing major challenges with regards to ILM policy for all my indexes.
In ILM policy I'm using two phase HOT and DELETE.
Hot Phase- It should have max 2mb of data or data should be there for 3 min then rollover.
Issue faced- The data keep on writing on same index, i.e became 5mb and after some time it got rollover i.e it didn't even follow 3min defined in policy.
Why dosent policy is followed strictly like if more than 2MB or more than 3 min either it should have rolled over.But it takes some random time to get rolled over in HOT face.
Delete phase- Defined to delete data after 2 seconds from rollover,even though now index was rollover as per above condition but didnt got deleted after 2 seconds.
What if the data become to large then this delay will also increase?,this is a major issue.
Ideally my data can be from 1GB to 30 Gb or 500Mb as well and want to delete this data after 1 month and new data should be rolled over to new index.But above is the challenge faced.
Hi @PRASHANT_MEHTA Those sizes and timeframes are too short..
ILM is not really design to operate on "Minute" Scheduled or Small KB/MB indices is designed to manage large scale indices GB, Hours, Days etc. and for that it works very well.
I posted a bit about it here and another elastician here
Hello @stephenb
Can you please tell me , what if I want to rollover my indices on monthly basis.
I want to generate rollover indices for SEPT...OCT...NOV...DEC
I know we can use "max_age":"30d" for that , but it will create problem for the months having 31 days & 28 days in a year.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.