elastic/elasticsearch

Add CBOR-friendly bulk format

Open

#8481 opened on Nov 14, 2014

View on GitHub
 (4 comments) (0 reactions) (1 assignee)Java (76,700 stars) (25,882 forks)batch import
:Distributed/CRUD>enhancementTeam:Distributedhelp wanted

Description

The current bulk format in JSON uses the \n as end-of-line markers, allowing the coordinating node to parse just the metadata line, then skip over the body line.

CBOR doesn't have any disallowed characters or character sequences, so this mechanism will not work. Instead we should have a CBOR specific format which looks like this:

{ metadata line }
int_length_of_body_line
{ body line }

The int_length_of_body_line would allow us to skip to the next metadata line.

For the delete action, no int_length_of_body_line would be required, as we already know that it can't have a body.

Depends on #7640

Contributor guide

Tech stack
java
Domain
backendapi
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
blocked
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
mostly clear
Prerequisites
Understanding of Elasticsearch bulk APIFamiliarity with CBOR encoding
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
20
Research direction
This issue depends on #7640, which should be resolved first. The current JSON bulk format uses newline delimiters, but CBOR has no disallowed characters, so a new format is needed: metadata line, integer length of body line, then body line. For delete actions, no length is required as they have no body. Implementation should modify the bulk endpoint to support this CBOR specific format. Consult the existing bulk request handling code in the server module, likely in org.elasticsearch.action.bulk.