make os thread nice values configurable#86703
Conversation
|
Workflow [PR], commit [345ce97] Summary: ❌
|
| MAKE_OBSOLETE(M, Bool, enable_variant_type, true) \ | ||
| MAKE_OBSOLETE(M, Bool, enable_dynamic_type, true) \ | ||
| MAKE_OBSOLETE(M, Bool, enable_json_type, true) \ | ||
| MAKE_OBSOLETE(M, Int64, os_thread_priority, 0) \ |
There was a problem hiding this comment.
@mstetsyuk let's make it alias to os_threads_nice_value_query otherwise it is a backward incompatible change
|
Do we have a profit from granular os nice values? Do we use them? Do we have tests (except that which only turn it ad check that it just not crash)? |
| else | ||
| res_group = create(context); | ||
| { | ||
| const Int32 os_threads_nice_value = context->getSettingsRef()[Setting::os_threads_nice_value_materialized_view]; |
There was a problem hiding this comment.
This almost newer happens. That else is needed only for corner cases ie Buffer flush.
It's basically a failed experiment - we don't use them for now |
|
But I don't rule out that, in the future, we may want to give more priority to some "niche" threads for performance tuning (e.g. in the keeper) |
Changelog category (leave one):
Changelog entry (a user-readable short description of the changes that goes into CHANGELOG.md):
Make nice values of Linux threads configurable to assign some threads (merge/mutate, query, materialized view, zookeeper client) higher or lower priorities.
Documentation entry for user-facing changes