there is an Elasticsearch database already built by Elrond team. In order to parse all the blocks from Elrond, we actually query from Elasticsearch which the block is already indexed. However, with "scroll" in Elasticsearch you can only query data from the latest to the oldest by scrolling. Also, only one worker can work on that because when I query with scroll, scroll_id for the next scroll is obtained only when the previous page is fetched. It is not possible to query scrolls from multiple workers.
Implementation proposal
One of the solutions is we have a single worker that fetches all the history blocks by scrolling, and another worker fetches the latest blocks at the same time. And there are stand-alone workers on fetching on accounts/ transactions etc. So the Worker would be useless and all this do is multiple threading fetching elasticsearch db into the juno db
Maybe there are better solutions but I don't know what is that atm
Feature description
there is an Elasticsearch database already built by Elrond team. In order to parse all the blocks from Elrond, we actually query from Elasticsearch which the block is already indexed. However, with "scroll" in Elasticsearch you can only query data from the latest to the oldest by scrolling. Also, only one worker can work on that because when I query with scroll, scroll_id for the next scroll is obtained only when the previous page is fetched. It is not possible to query scrolls from multiple workers.
Implementation proposal
One of the solutions is we have a single worker that fetches all the history blocks by scrolling, and another worker fetches the latest blocks at the same time. And there are stand-alone workers on fetching on accounts/ transactions etc. So the Worker would be useless and all this do is multiple threading fetching elasticsearch db into the juno db
Maybe there are better solutions but I don't know what is that atm