Refactor env vars to clarify they are for individual ranks #449
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Changes included in this PR
See here for more context. Essentially the environment variables for the buffer pool are split between cluster-level, node-level and rank-level granularity, which can cause confusion.
For example:
BODO_BUFFER_POOL_MEMORY_SIZE_MiBspecifies the memory available to this rank for the buffer pool, whereasBODO_BUFFER_POOL_STORAGE_CONFIG_1_SPACE_PER_DRIVE_GiBIs the space available for each node (assuming one drive).Ideally IMO
BODO_BUFFER_POOL_MEMORY_SIZE_MiBshould just specify the amount of space available to all processes (on the same node) and then that space gets divided among ranks evenly, but then it creates a weird situation withBODO_BUFFER_POOL_MALLOC_FREE_TRIM_THRESHOLD_MiBstill being per rank.Adding "worker" makes it a bit more clear but then the environment variables too long e.g.
BODO_BUFFER_POOL_WORKER_MALLOC_FREE_TRIM_THRESHOLD_MiBTesting strategy
User facing changes
Checklist
[run CI]in your commit message.