esbulk 0.4.9

• add man page

esbulk 0.4.8

• more backwards compat fixes

esbulk 0.4.7

• 1.4.X compat fixes

esbulk 0.4.6

• try to handle case, where creating an already existing index results in a HTTP 400 (e.g. elasticsearch 1.4.3)

esbulk 0.4.5

• fixes #10

esbulk 0.4.4

• API adjustements and fixes: #9

esbulk 0.4.3

• add support for nested ids and concatenated ids via #8, thanks @sakshambathla

esbulk 0.4.2

• add X-Pack, HTTP basic auth support, with curl syntax:

$ esbulk -u username:password -index hi file.ldj

esbulk 0.4.1

Previously it was possible to silently lose documents, if a bulk operation failed. Starting with this release, the bulk operation error is checked and the indexing is aborted, in case the operation failed.

There are two ways to deal with such an error:

• reduce the number of workers (using the defaults should be fine)
• increase thread_pool.bulk.queue_size in your elasticsearch settings

This severe error occured in scenarios, where esbulk would exhaust the thread pool quickly (e.g. high number of workers and small elasticsearch thread pool queue size). Using the defaults (e.g. number of workers equals number of cpus) would result in correct results in most cases, which is probably why this bug could linger so long.

esbulk 0.4.0

• attempted fix for #5