I'm doing a backup in our S3. So far we're using a monthly index, but we're planning to do a daily backup using curator.
Versions:
Elasticsearch 5.6.1
Curator 5.2
Here are some questions:
Is it good to do a daily snapshot for each index ?
Using the curator filters, is it possible to grab the current date? because when using timestring and regex, it still grabs all the similar pattern which is index-YEAR-MONTH.
For snapshot, is it okay to include old indices even it will be deleted since I'm going to create snapshot based from creation time
Since we're using a monthly index, we expect to create a snapshot while there's an on-going index. Based from the docs, this shouldn't be an issue.
Is it good to do a daily snapshot for each index ?
On cloud.elastic.co we are doing snapshot every 30 minutes. Everyday is fine.
Using the curator filters, is it possible to grab the current date? because when using timestring and regex, it still grabs all the similar pattern which is index-YEAR-MONTH.
Since I only need around 3 months worth of data, I plan to delete indices which are older than 3 months. Let's say it's already November, the index-2017-07 will be deleted. Will that affect the snapshot created when index-2017-07 is present? Based from the docs, it shouldn't since it's incremental.
For the curator part, let's say I have index named "filebeat-%Y-%m". So far, I have indices for filebeat starting June which results to indices below:
filebeat-2017-06
filebeat-2017-07
filebeat-2017-08
filebeat-2017-09
filebeat-2017-10
Right now it's October, Is there a way in curator to just grab the current month which is October and do a snapshot of it? or that's a not good thing since if I'm doing a backup daily then it doesn't matter if I include all filebeat-%Y-%m indices since it's incremental.
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.