Automatic management of indexing buffer size #241
Comments
|
Automatic management of indexing buffer size, closed by fdb2eff. |
mindw
pushed a commit
to mindw/elasticsearch
that referenced
this issue
Sep 5, 2022
whiteboard nginx * Merged in nginx security headers for web apps (pull request elastic#229) * Fix MPC-2824: disallowed loading student and admin portal in iframe * Add headers to nginx webapp and fabfile * X-Frame-Options: deny clickjack prevention telling browser not to load our site in iframe * X-Content-Type-Options: nosniff prevent browser from interpreting content type * X-Permitted-Cross-Domain-Policies: none prevent cross-domain loading in adobe products * Referrer-Policy: no-referrer prevent browser from including referrer header * Permissions-Policy: -list- permission list for the site (previously Feature-Policy) * Strict-Transport-Security: -list- prevent browser from protocol downgrade and cookie hijacking * Remove nginx version from server header Approved-by: Matan Yemini Approved-by: Can Yildiz Approved-by: Gideon Avida Approved-by: Andre Sodermans * 1.11.0-915-dev * Merged in dev/ttang2/dequeue_MPC-2823_21.09.08 (pull request elastic#228) set max_group_prepared_capacity to min_size * set max_group_prepared_capacity to min_size * add comments * set warm pool max size and not set desired number * Merged master into dev/ttang2/dequeue_MPC-2823_21.09.08 * Merged master into dev/ttang2/dequeue_MPC-2823_21.09.08 * Merged master into dev/ttang2/dequeue_MPC-2823_21.09.08 * remove extra for wram pool max size * Merge branch 'dev/ttang2/dequeue_MPC-2823_21.09.08' of bitbucket.org:engageli/devops into dev/ttang2/dequeue_MPC-2823_21.09.08 * Merged master into dev/ttang2/dequeue_MPC-2823_21.09.08 * Merge branch 'dev/ttang2/dequeue_MPC-2823_21.09.08' of bitbucket.org:engageli/devops into dev/ttang2/dequeue_MPC-2823_21.09.08 Approved-by: Gideon Avida * Merged in dev/can/fix-ami-retrieval-master (pull request elastic#235) Merged in dev/can/fix-ami-retrieval-2021-09-15 (pull request elastic#234) Fix ami retrieval for merger and recorder Also update use1a stack to 1.10.23 Approved-by: Gideon Avida * Merged in dev/can/MPC-2843-include-u-rooms-2021-09-09 (pull request elastic#236) MPC-2843: Include u rooms in attendance report With the caveat of missing tenant names for now * Merged in dev/can/attendance-improvements-2021-09-16 (pull request elastic#237) Add date to report name * Merged in dev/can/coventry-stacks-ansible-2021-09-14 (pull request elastic#238) Add coventry stacks to ansible hosts * 1.11.1-917-dev * Small change to optionally print user aliases from google directory * whiteboard nginx Approved-by: Gideon Avida
emilykmarx
pushed a commit
to emilykmarx/elasticsearch
that referenced
this issue
Dec 26, 2023
This issue was closed.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Currently, there is an option to set the
ram_buffer_sizeon the robin engine, but this is set on a per shard level. So, it can be set to64mb, but it does not take into account the number fo total shards allocated on the nodes. For example, if there is a single shard, and we have plenty of memory, then we should set it to a higher value, and if there are a lot of shards, this should be set to a lower value.A global management of the indexing buffer size should be maintained, with the ability to set the memory used for indexing as either a value or a percent of the allocated memory to the jvm, across all shards.
The setting is called
indices.memory.index_buffer_size, and it defaults to40%. It can either be set to a percent value (with%suffix), or a byte size value (for example1.5gb).The default of the setting is
40%.The text was updated successfully, but these errors were encountered: