This is a community forum, so please be patient. It may take a few days to get a response, especially if your question is very specific and only a few people may know the answer.
While I do not know the answer, it sounds to me like a theoretical rather than practical question. If you assume rollover happens every hour, which is quite frequent, it would take 114 years to reach the maximum value. If you are rolling over indices much more frequently than that I would recommend you look at how you handle sharding as that to me would sound very suboptimal.
The whole point with rollover is to be able to achieve a suitable shard size in the face of fluctuating ingest volumes. You should avoid creating a lot of very small shards as this adds overhead and can cause performance and stability issues.
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.