The HBase read path
When a read for a key is issued from the HBase client, as with writes, the client needs to query the META table to identify the RegionServer that contains the data for the given key.
Once the RegionServer receives the read request, it looks it up in the Memstore. However, this by itself is insufficient. The contents of the Memstore might have been flushed to disk by the time the read request arrives, so the RegionServer has to look for the key in the HFile that the Memstore contents were previously flushed into. However, it's not sufficient to look at the most recent HFile for the region since writes for that key could have arrived at any time in the past, so the RegionServer has to look for the key in every HFile that ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access